IN RE: UBER TECHNOLOGIES (2024)

PRETRIAL ORDER NO. 17: ORDER ON MOTIONS TO DISMISS PURSUANT TO CALIFORNIA AND TEXAS LAW

I. BACKGROUND

A. Procedural Background

On October 4, 2023, the Judicial Panel on Multidistrict Litigation (“JPML” or “the Panel”) created MDL No. 3084, centralizing 22 actions in this Court for coordinated pretrial proceedings. In re Uber Techs., Inc., Passenger Sexual Assault Litig., 699 F.Supp.3d 1396 (Jud. Pan. Mult. Lit. 2023) (“Transfer Order”).

The Court established a case management plan under which Plaintiffs were to file a Master Long-Form Complaint (“MC”) and Uber would file motions to dismiss under the laws of five states of its choosing. Uber timely filed motions under the laws of California, Texas, Illinois, New York, and Florida. This order addresses the motions filed under the laws of California and Texas.

B. Allegations in the Master Complaint

Beginning in 2013, Uber (following shortly on the heels of its competitor, Lyft) started offering a novel form of app-enabled transportation, where Uber would connect people who needed rides with individuals who had a driver's license, access to a car, and willingness to drive. MC ¶¶ 5–7. In doing so, Uber helped to overturn the existing paradigm in the ride industry. This, according to Plaintiffs, required grappling with the existing public perception that it would have been “unreasonably dangerous for a young woman (or any solo traveler for that matter) to enter a complete stranger's car in order to reach her destination.” Id. ¶ 1. It was understood that a foreseeable risk of hitchhiking was sexual assault. Id. Plaintiffs say that it was considered reasonably safe to hail a licensed taxi, which were strictly regulated, including regarding who could become a taxi driver. Id. ¶ 2.

The MC alleges that Uber nevertheless understood that the impression of safety was important: to make money, Uber had to induce “people to voluntarily enter into an enclosed environment controlled by a random stranger,” and to do so it knew that it had to “cultivate a reputation for safety and transfer that reputation to Uber drivers.” Id. ¶¶ 189, 193.

While a reasonably careful enterprise might “adopt a ‘think before you act’ way of doing business,” Uber allegedly ignored the foreseeable risk of sexual assault its platform created for most of its early history. Id. ¶ 112. Instead, it “chose rapid growth at all costs.” Id. ¶ 10. It did not hire safety experts, study or plan for the risk of sexual assault, devote staff or resources to mitigating or understanding that risk, or design its app or business model in a way that would avoid sexual assault; in short, according to Plaintiffs, Uber did not “spend a single minute thinking about how to prevent sexual assault” on its platform. Id. ¶¶ 16, 125–28. Instead, it adopted a strategy designed to “make the acquisition of drivers and riders as cheap and frictionless as possible” in order to grow the platform as rapidly as possible. Id. ¶ 11. By 2014, less than a year after Uber started onboarding non-professional drivers, it started receiving reports of sexual assault committed by Uber drivers against riders. Id. ¶ 258. It has continued receiving these reports from 2014 to the present. Id.

By 2017, Uber began to appoint employees to positions ostensibly focused on sexual violence prevention, although it did not require any expertise or background in safety fields. Id. ¶ 129. Instead, according to Plaintiffs, those employees were primarily focused on managing the problem in a way that would protect Uber's reputation, rather than in a way that would protect passengers. Id. In 2021, the New York Times reported that Uber's team of purported “investigators” typically had “little to no background in trauma response ․ They are given scripts to read to often-upset passengers and drivers, and strict company policies prohibit the agents from reporting the incidents to the police, even when perpetrators acknowledge their actions.” Id. ¶ 256. To this day, Uber prohibits its agents from reporting sexual assault incidents to police. Id.

1. Uber's Control Over its Platform

The MC alleges that Uber has “unparalleled knowledge and comprehensive control with respect to every aspect of its transportation model, including those that determine whether rides are safe or unsafe.” Id. ¶¶ 70, 122, 130–34. Uber allegedly controls the design of the app, where rides begin and end, the route they take, the matching of drivers with passengers, what information passengers receive about drivers, the terms of Uber's relationships with its drivers, and the conditions of those drivers’ work. Id. ¶¶ 51–111. Uber handles all complaints and grievances regarding rides. Id. ¶ 68. Uber tracks drivers and riders using GPS, and it has the capability to intervene in real time when it detects unusual activity during a ride. Id. ¶¶ 92, 103–07.

2. Uber's Safety-Related Marketing

Part of Plaintiffs’ theory is that Uber has long marketed itself in a misleading way, particularly when it comes to safety. Plaintiffs allege that Uber understands “that passengers are largely motivated by perceptions of safety when choosing Uber's services.” Id. ¶ 13. And they allege that to get its business model off the ground, Uber understood that it had to manage potential riders’ perceptions about the safety of getting into a stranger's car alone. Id. ¶ 12.

Accordingly, Uber worked to assure passengers that it was offering what its marketing called the “safest rides on the road,” “a ride you can trust,” and a service that was “created to ensure reliable access to safe rides.” Id. Part of Uber's marketing also concerned its relationships with drivers. Plaintiffs point out that Uber's marketing tends to hold out its rides not as independent transactions between private citizens merely brokered by Uber, but as rides offered by Uber. In other words, Uber allegedly held itself out as a transportation company, not merely a technology company, and it held the drivers out as its drivers, not independent people with cars. See, e.g., ¶¶ 196, 199, 12. Uber lends “the imprimatur of a legitimate corporation” to its drivers. Id. ¶ 13. For example, Uber provides its drivers with Uber decals and signs to adhere to their cars. When it does so, according to Plaintiffs, it “intends the Uber decal and other matching features to reassure the prospective passenger: ‘This is not just a random stranger's car. It is an UberX and it is safe to get in.’ ” Id. ¶¶ 191–93.

Uber's initial advertising about safety was, allegedly, brazenly false. In the early years of UberX (Uber's main rideshare business), the company advertised that it offered “the safest ride on the road,” that it sets “the strictest standards possible”, and that it “aims to go above and beyond local requirements” with “gold standard” and “industry leading” background checks. Id. ¶¶ 20, 206-08. Beginning in 2014, Uber charged a “Safe Ride Fee” for every ride, which it told the public “support[ed] our continued efforts to ensure the safest possible platform for Uber riders.” Id. ¶ 203. In reality, the “Fee” was never earmarked for safety but was just profit margin. Id. ¶ 204. Uber eventually paid more than $50 million to settle lawsuits based on such marketing, and it changed its safety representations to be somewhat less aggressive, although they still emphasized Uber's supposed commitment to safety. Id. ¶¶ 20, 206-08. For example, Uber continued to represent that “At Uber, Safety Never Stops,” that it offered “Safe, reliable rides in minutes,” that users could “Ride with confidence—The Uber experience was built with safety in mind,” and to say that Uber is “Designing a safer ride” and has “features to keep you safe.” Id. ¶¶ 20, 211–23.

The MC also describes advertising campaigns specifically directed to women and intoxicated riders (and intoxicated women riders), whom it says Uber “disproportionately targets” for advertising. For example, some ad campaigns describe how Uber is “driving women's safety forward,” and its “marketing photos and videos predominantly feature smiling women riding in Ubers.” Id. ¶¶ 21, 225. Until recently, its website stated it was “committed to help stop incidents before they happen.” Id. Uber also allegedly markets its rides as a safe, smart transportation option for intoxicated riders, including through partnerships with Mothers Against Drunk Driving and with alcohol manufacturers and bars. Id. ¶¶ 22, 233–37.

3. Actual and Possible Safety Measures

The complaint alleges that, because Uber needed as large a supply of drivers as possible (in order to make sure rides were widely available and arrived quickly), it was incentivized to avoid imposing safety measures that would deter drivers. Accordingly, Uber does not interview drivers, meet with them, require any prior experience, solicit references, do any drug or alcohol testing, or require any tests. Id. ¶¶ 147, 167–68.

Instead, Uber uses only a cursory background check, for which it employs a background check company that promises to vet drivers within 36 hours; sometimes Uber enrolls drivers before the check is even complete. Id. ¶¶ 17, 145, 165. These background checks are name-based, with relatively short coverage periods, and they are apparently known to have fundamental flaws and be vulnerable to fraud. Id. ¶¶ 143–44. When some states performed their own fingerprint-based background checks, 12-15% of drivers who were eligible under Uber's standards flunked. Id. ¶¶ 17, 152–53. The background checks are allegedly inferior to those used by the taxi industry, which Plaintiffs say generally uses fingerprinting or other biometric data and/or runs applicants against the best possible databases. Id. ¶¶ 17, 142.

Plaintiffs allege that Uber had many options for more rigorously and seriously screening drivers. For example, it could have implemented:

rigorous screening of drivers; prompt termination of drivers who engage in sexual misconduct; in-App tools such as panic buttons with prompt responses from Uber and coordination with law enforcement; surveillance for unusual conduct (such as lengthy stops) during rides; training drivers and communicating to them that there will be strict penalties and accountability; diligently investigating allegations of sexual assault; communicating to riders that because Uber's drivers are not professionals and are not well known to Uber, Uber is unsafe and unsuitable for use by solo intoxicated riders; immediate termination of drivers who violate policies, [or] mandatory installation of cameras that could not be deactivated during trips[.]

Id. ¶ 131. It also allegedly has long known that requiring in-ride cameras would deter assault, but has not done so because it would “slow onboarding of new drivers, discourage some drivers from signing up ․, and jeopardize Uber's argument that drivers are independent contractors.” Id. ¶ 25. Uber also allegedly “does not adequately monitor drivers’ safety performance” and “makes it hard for riders to report sexual assault.” Id. ¶ 26. Plaintiffs suggest that this practice deprives Uber of “the true data on sexual assault,” resulting in artificially low rates of reporting.

Uber also allegedly does not handle complaints in a manner designed to protect riders. Until at least 2019, “Uber's official policy has been to require more than one sexual assault complaint before terminating a driver. Depending on how much money a particular driver was making for Uber, Uber would sometimes tolerate three or four sexual assaults before terminating the driver. It is unknown how many of these drivers, who received ‘free passes’ for one, two, or three sexual assaults, are still driving for Uber.” Id. ¶¶ 28, 297. Uber also has a policy of not reporting drivers’ criminal sexual assaults to law enforcement, even if the driver admits the crime to Uber. Id. ¶¶ 304–05. As a result, Plaintiffs say, driving for Uber is a near-ideal environment for sexual predators.

4. The Uber App

Passengers and drivers access Uber rides through a mobile app that they must install on their phones. Id. ¶¶ 446–47. The MC describes the Uber app essentially to include the totality of Uber's technological platform, including the front-end user interface (the thing that an individual user must download and install onto her phone) and the back-end infrastructure, including Uber's algorithms and nationwide network of servers. So defined, the MC alleges that the app controls every aspect of one's transportation through Uber, including matching drivers and passengers, telling the driver what route to take, setting the terms of payment, fielding feedback and complaints, and gathering massive amounts of data for Uber. Id. ¶¶ 60–71, 90–94, 58. Accordingly, Plaintiffs’ argue that “[t]he driver is merely a component of Uber's system,” and “Uber alone[ ] decides when and how to use its data and its App to shape the transportation experience.” Plaintiffs’ theory is that the Uber app is a product for purposes of products liability law.

The app has some safety features, including GPS tracking and an emergency button that dials 911, but Plaintiffs allege that these are essentially for show, and that many better design options are available. Id. ¶¶ 24, 473, 487, 497. For example, Uber uses GPS tracking to match drivers with nearby riders, but it does not use the feature to alert the police when there is a suspicious route deviation. Id. ¶¶ 65, 473, 487, 497. Plaintiffs also say that Uber could have, but did not, code its App to take advantage of the capabilities of most smartphones, such as biometric reading and video recording. Id. ¶¶ 473, 487.

5. Plaintiffs’ Claims and Uber's Motions to Dismiss

The MC alleges the following eight claims (which, for whatever reason, are listed alphabetically beginning with “B”):

• Claim B: Negligence (including Negligent Hiring, Retention, Supervision, and Entrustment)

• Claim C: Fraud and Misrepresentation

• Claim D: Negligent Infliction of Emotional Distress

• Claim E: Common Carrier's Non-Delegable Duty to Provide Safe Transportation (Certain States Only)

• Claim F: Other Non-Delegable Duties to Provide Safe Transportation (Certain States Only)

• Claim G: Vicarious Liability for Drivers’ Torts (Employee, Retained Control, Apparent Agency, Ratification, California Public Utilities Code)

• Claim H: Strict Products Liability (Failure to Warn and Design Defect)

• Claim I: California Unfair Competition Law

Plaintiffs seek damages, punitive damages, and “appropriate declaratory, injunctive, and equitable relief.”

In its California Motion to Dismiss (dkt. 384), Uber moves to dismiss all of Plaintiffs’ claims except Claims B and E. In its Texas Motion to Dismiss (dkt. 387), Uber moves to dismiss all of Plaintiffs’ claims except Claim B. It also argues in both motions that Plaintiffs have failed to plead facts that would entitle them to punitive damages and that Plaintiffs cannot seek injunctive relief.

II. LEGAL STANDARD

Under Federal Rule of Civil Procedure 12(b)(6), a district court must dismiss a complaint if it fails to state a claim upon which relief can be granted. To survive a Rule 12(b)(6) motion to dismiss, the plaintiff must allege “enough facts to state a claim to relief that is plausible on its face.” Bell Atl. Corp. v. Twombly, 550 U.S. 544, 570, 127 S.Ct. 1955, 167 L.Ed.2d 929 (2007). A claim is facially plausible when the plaintiff pleads facts that “allow the court to draw the reasonable inference that the defendant is liable for the misconduct alleged.” Ashcroft v. Iqbal, 556 U.S. 662, 678, 129 S.Ct. 1937, 173 L.Ed.2d 868 (2009) (citation omitted). There must be “more than a sheer possibility that a defendant has acted unlawfully.” Id.

While courts do not require “heightened fact pleading of specifics,” a plaintiff must allege facts sufficient to “raise a right to relief above the speculative level.” Twombly, 550 U.S. at 555, 570, 127 S.Ct. 1955. In deciding whether the plaintiff has stated a claim upon which relief can be granted, the Court accepts the plaintiff's allegations as true and draws all reasonable inferences in favor of the plaintiff. See Usher v. City of Los Angeles, 828 F.2d 556, 561 (9th Cir. 1987). But the court is not required to accept as true “allegations that are merely conclusory, unwarranted deductions of fact, or unreasonable inferences.” In re Gilead Scis. Sec. Litig., 536 F.3d 1049, 1055 (9th Cir. 2008).

III. DISCUSSION

A. Pleading Issues

Before proceeding to particular claims, some comments about the pleadings—and the applicable pleading standard—are in order. Here, as in many other MDLs, the plaintiffs have handled their pleadings through a master complaint and individual short-form complaints. Plaintiffs do not intend the MC to supersede any individual pleadings, but instead ask that it be regarded as an administrative device designed to summarize allegations and claims that are common to the cases in this MDL. See MC at 1. Accordingly, the MC primarily makes allegations about Uber's conduct that are relevant to plaintiffs’ various theories of Uber's liability for their assaults by drivers. The MC does not have many case-specific allegations. For example, it does not allege when, where, or how plaintiffs were assaulted or harassed; it does not allege what sorts of marketing individual plaintiffs saw or relied on; and it does not allege how Uber handled complaints about specific drivers’ conduct. The MC is honest about this: it suggests that “Plaintiff-specific fact[s]” will be alleged in underlying complaints or short-form complaints. Id. All of that is quite conventional. See, e.g., Gelboim v. Bank of Am. Corp., 574 U.S. 405, 413 n.3, 135 S.Ct. 897, 190 L.Ed.2d 789 (2015) (mentioning MDL procedure by which “the master complaint is not meant to be a pleading with legal effect but only an administrative summary of the claims brought by all the plaintiffs”) (quoting In re Refrigerant Compressors Antitrust Litigation, 731 F.3d 586, 590 (6th Cir. 2013)).

But a master complaint procedure—at least the kind that the parties and the Court contemplated—does not excuse Plaintiffs from the obligation to plead the facts that are necessary to support their claims. Here, Uber has challenged some of Plaintiffs’ claims on the grounds that they are not viable in the absence of certain individualized allegations. This is not the basis of most of Uber's arguments, but the issue does recur in a few different places throughout the complaint. In general, Plaintiffs’ response to these arguments is that a master complaint is merely an administrative device, and the time for dealing with individualized questions is later. The opposition brief does not cite or refer to any individual allegations in the short-form complaints (although Plaintiffs filed their short-form complaints after Uber's motions to dismiss were filed and before the opposition was due). Perhaps more troublingly, it seems like Plaintiffs could not have pointed to individual allegations even if they wanted to: on a spot-check of short-form complaints, the Court found none that actually pled the additional Plaintiff-specific facts that are envisioned by the MC. In effect, that means that the operative pleading in any given case consists solely of the allegations in the MC.

Under these circumstances, it makes little sense for the Court to defer decision on Uber's arguments that relate to the lack of individualized allegations. It would be one thing if the individualized allegations were present in any individual pleadings, but they do not seem to be. Although these cases have been centralized in an MDL, Plaintiffs have filed individual cases that have to be held to the usual pleading standards. See, e.g., In re Korean Air Lines Co., Ltd., 642 F.3d 685, 700 (9th Cir. 2011) (“There is much, of course, that an MDL court can do in its sound discretion in order to manage multidistrict litigation effectively. It can designate a lead counsel. It can hold some cases in abeyance while proceeding with others. In discretionary matters going to the phasing, timing, and coordination of the cases, the power of the MDL court is at its peak. But when it comes to motions that can spell the life or death of a case, such as motions for summary judgment, motions to dismiss claims, or, as here, a motion to amend pleadings, it is important for the district court to articulate and apply the traditional standards governing such motions.”). In other words, Uber is entitled to test the legal sufficiency of Plaintiffs’ claims by way of Rule 12, just as it would be in any other litigation. See id. As things stand, the absence of case-specific allegations will require the dismissal of certain claims.¹ The Court cannot deny Uber's motion on the grounds that some plaintiffs might be able to plead such facts later.² Nor can the Court rule that, hypothetically, a plaintiff would plausibly allege a given claim if she alleged certain other facts. That is not how Rule 12 works.

In a recent case management conference, Plaintiffs proposed a mechanism for adjudicating individualized pleading issues where the parties would select a handful of cases in which to file long-form pleadings with more substantial allegations. Then Uber would challenge those pleadings, and the parties could discuss how the Court's rulings applied to other cases. A version of this practice could well make sense later, but no one has ever proposed it before in this MDL. The time for devising a case management plan for adjudicating Rule 12 motions is not after hundreds of pages of briefing have been completed on Rule 12 motions. At this juncture, Plaintiffs’ pleadings will stand or fall according to the ordinary pleading standards—which is to say, according to the allegations that the plaintiffs have actually made. Where the pleadings are deficient but might be curable with additional allegations, the Court will grant leave to amend. At that point, the parties and the Court can discuss the best way to manage the filing of amended pleadings and any challenges to them.

B. Vicarious Liability Theories

Plaintiffs plead four theories of vicarious liability. The first, which is brought under California but not Texas law, is the traditional doctrine that employers are vicariously liable for the torts of their employees when those torts are within the scope of employment. The second is an “apparent agency” theory. Under that doctrine, a principal can be liable for the torts of someone who he holds out as his agent if the plaintiff relies on those representations and is injured as a result of his reliance. The third is a ratification theory. There, the idea is that a principal can be liable for the torts of an agent, even if they were outside the scope of the agency relationship, if the principal subsequently ratifies the tort. The fourth theory is based on Uber's non-delegable duty of care as a common carrier. The basic idea is that common carriers have non-delegable duties to safely transport their passengers, so any intentional torts committed by the agents of a common carrier against a passenger can create direct liability for the carrier itself—essentially, a form of vicarious liability.

At this stage, Uber does not argue that its drivers are not its employees (or otherwise its agents). Instead, it argues that even if Uber's drivers were its employees, Uber still could not be vicariously liable for their alleged assaults under any of Plaintiffs’ theories. This section begins by addressing Uber's threshold argument that a Texas statute bars vicarious liability of any sort for companies like Uber. Then, it addresses each of the theories of vicarious liability in turn.

1. Texas's TNC Statute

Last year, the Texas legislature enacted Tex. Civ. Prac. & Rem. Code § 150E.002-003, which limits the circumstances under which a “transportation network company” (“TNC”) can be held vicariously liable for damages. The statute is effective as of September 1, 2023. It applies to actions where:

(1) a transportation network company is a defendant;

(2) the claimant seeks recovery of damages for loss of property, bodily injury, or death;

(3) the claim for which the action or proceeding is brought arises out of the ownership, use, operation, or possession of a network vehicle while the vehicle's driver or passenger was logged on to a transportation network company's digital network; and

(4) the theory of recovery for which damages are sought against the transportation network company is based on:

(A) the ownership, operation, design, manufacture, or maintenance of a digital network accessed by a driver or passenger; or

(B) the relationship, affiliation, or interaction with a driver logged on to a transportation network company's digital network.

Tex. Civ. Prac. & Rem. Code § 150E.002(a). Where these criteria are met, a TNC “may not be held vicariously liable for damages ․ if:”

(1) the claimant does not prove by clear and convincing evidence that the company was grossly negligent with respect to the subject claim; and

(2) the company has fulfilled all of the company's obligations with respect to the transportation network company driver under Chapter 2402, Occupations Code, relating to the subject claim.

Id. § 150E.003(a).

Uber argues that this statute, which the Court will call the “Texas TNC Statute,” precludes it from being subject to vicarious liability under Texas law for sexual assaults by its drivers. No one disputes that Uber is a “transportation network company” under the meaning of the statute. But Plaintiffs argue that the statute should be read to apply only to people injured by vehicles in auto accidents. Since the statute was passed relatively recently, there do not seem to be any Texas cases construing it yet.

Deciding how the TNC applies to claims for sexual assaults by rideshare drivers is an exercise in statutory interpretation. In Texas, a court deciding the meaning of a statute must “ascertain and give effect to the Legislature's intent.” Entergy Gulf States, Inc. v. Summers, 282 S.W.3d 433, 437 (Tex. 2009) (citing F.F.P. Operating Partners., L.P. v. Duenez, 237 S.W.3d 680, 683 (Tex. 2007)). “Where text is clear, text is determinative of that intent.” Id. (citing State v. Shumake, 199 S.W.3d 279, 284 (Tex. 2006)). Only when the legislature's “words are ambiguous do we ‘resort to rules of construction or extrinsic aids.’ ” Id. (quoting In re Estate of Nash, 220 S.W.3d 914, 917 (Tex. 2007)).

Applying these principles leads to the conclusion that the Texas TNC Statute bars vicariously liability under Texas law for claims like Plaintiffs’. This much follows from the unambiguous language of the statute. Notwithstanding Plaintiffs’ arguments, their claims “seek recovery for damages for ․ bodily injury”; they uniformly “arise out of the ownership, use, operation, or possession of a network vehicle while the vehicle's driver or passenger was logged on to a transportation network company's digital network”; and the “theory of recovery is based on ․ the relationship, affiliation, or interaction with a driver logged on to a transportation network company's digital network.” Tex. Civ. Prac. & Rem. Code § 150E.002(a). It is not possible to read the statute as limited to claims arising out of car accidents without reading in language that simply is not there. In other words, if the Texas Legislature wanted to limit the law's applicability to auto accidents, it easily could have done so by expressly limiting the types of claims to which the statute would apply. Instead, it used broad language that captures claims against TNCs based on injuries to plaintiffs caused by TNC drivers while those drivers are driving for the TNC.

The structure of the statute confirms that it reaches sexual assault claims. For its vicarious liability limitation to apply, the TNC must (1) not be grossly negligent and (2) be in compliance with its “obligations with respect to the transportation network company driver under Chapter 2402, Occupations Code, relating to the subject claim.” Id. § 150E.003(a). Those “obligations” under the Occupations Code include the conduct of a “local, state, and national criminal background check” for the driver that “includes the use of ․ (B) the national sex offender public website maintained by the United States Department of Justice or a successor agency[.]” Tex. Occ. Code § 2402.107(a)(2). Conditioning the statute's protections on compliance with these obligations would make little sense if the statute were only applicable to car accidents. Instead, the idea seems to be to substitute the Occupations Code's regulation of TNCs and their drivers for traditional doctrines of vicarious liability, to whatever extent they would apply to sexual assault or harassment claims. The text is clear, so there is no need to look to the legislative history that Plaintiffs cite. See Summers, 282 S.W.3d at 437.

Plaintiffs also suggest that their negligence allegations are sufficient to show that Uber was grossly negligent with respect to their claims, and that the TNC statute does not apply as a result. See Tex. Civ. Prac. & Rem. Code § 150E.003(a)(1). In Texas, gross negligence “means an act or omission”:

(A) which when viewed objectively from the standpoint of the actor at the time of its occurrence involves an extreme degree of risk, considering the probability and magnitude of the potential harm to others; and

(B) of which the actor has actual, subjective awareness of the risk involved, but nevertheless proceeds with conscious indifference to the rights, safety, or welfare of others.

Tex. Civ. Prac. & Rem. Code § 41.001(11); see also Tarrant Cnty. v. Morales, 207 S.W.3d 870, 874 (Tex. App.—Fort Worth 2006). This argument fares better. As discussed below in connection with the question of punitive damages, Plaintiffs’ allegations are sufficient to establish the elements of gross negligence.³ So if the facts ultimately support the view that Uber acted with gross negligence under Texas law, then Plaintiffs may be able get around the Texas TNC Statute's restriction on vicarious liability. Otherwise, any vicarious liability claims based on incidents occurring after September 1, 2023, will be barred under Texas law.⁴

2. Respondeat Superior (California Only)

Plaintiffs plead respondeat superior liability under the laws of California but not Texas. Opp'n at 40. Uber's California motion does not argue about whether there is an employer-employee or principal-agent relationship between Uber and its drivers. Instead, it argues that, even if such a relationship existed, there could be no respondeat superior liability for the alleged assaults.

a. Legal Standard

“The rule of respondeat superior is familiar and simply stated: an employer is vicariously liable for the torts of its employees committed within the scope of the employment.” Lisa M. v. Henry Mayo Newhall Meml. Hosp., 12 Cal.4th 291, 48 Cal.Rptr.2d 510, 907 P.2d 358, 360 (1995). But California does not follow the traditional common law rule that intentional torts can only be considered “within the scope of employment” where the employee acts to serve the employer's interests. See id., 48 Cal.Rptr.2d 510, 907 P.2d at 361 (discussing Carr v. Wm. C. Crowell Co., 28 Cal.2d 652, 171 P.2d 5 (1946)). Instead, California courts have formulated rules tailored to the underlying purposes of the respondeat superior doctrine. Its view is that “ ‘the modern justification for vicarious liability is a rule of policy, a deliberate allocation of a risk.’ ” John R. v. Oakland Unified Sch. Dist., 48 Cal.3d 438, 256 Cal.Rptr. 766, 769 P.2d 948, 955 (1989) (in bank) (quoting Hinman v. Westinghouse Elec. Co., 2 Cal.3d 956, 88 Cal.Rptr. 188, 471 P.2d 988, 990 (1970) (in bank)). In other words, under this version of respondeat superior, “ ‘[t]he losses caused by the torts of employees, which as a practical matter are sure to occur in the conduct of the employer's enterprise, are placed upon that enterprise itself, as a required cost of doing business.’ ” Id.; accord Farmers Ins. Group v. County of Santa Clara, 11 Cal.4th 992, 47 Cal.Rptr.2d 478, 906 P.2d 440, 448 (1995). The Supreme Court of California has described “three reasons for applying the doctrine of respondeat superior: (1) to prevent recurrence of the tortious conduct; (2) to give greater assurance of compensation for the victim; and (3) to ensure that the victim's losses will be equitably borne by those who benefit from the enterprise that gave rise to the injury.” Mary M. v. City of Los Angeles, 54 Cal.3d 202, 285 Cal.Rptr. 99, 814 P.2d 1341, 1343 (1991) (in bank); see also Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 366 (discussing these “three identified policy goals of the respondeat superior doctrine”).

Intentional torts can fall within the scope of this rule. See Mary M. v. City of Los Angeles, 285 Cal.Rptr. 99, 814 P.2d at 1344 (1991) (“Tortious conduct that violates an employee's official duties or disregards the employer's express orders may nonetheless be within the scope of employment.”). For example, in the seminal case of Carr v. Wm. C. Crowell Co., a construction worker threw a hammer at the head of the plaintiff—another worker—seriously injuring him. In holding that the assault was within the scope of employment, the Supreme Court of California reasoned that injury was “an outgrowth of [the assailant]’s employment” because “[n]ot only did the altercation leading to the injury arise solely over the performance of [the assailant]’s duties, but his entire association with plaintiff arose out of his employment on the building under construction. He had never seen plaintiff before the day preceding the accident, and had never conversed with him before the dispute over the plate.” Carr, 171 P.2d at 8. Sexual torts may also give rise to respondeat superior liability under at least some circumstances. The Supreme Court of California has expressly declined to follow the path of other jurisdictions that declare such torts per se beyond the scope of employment. See Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 363 (rejecting as “not ․ compelling” an argument that “sex crimes are never foreseeable outgrowths of employment because they, unlike instances of nonsexual violence, are not the product of ‘normal human traits’ ”); see also id., 48 Cal.Rptr.2d 510, 907 P.2d at 364.

While these basic standards are beyond dispute, the parties disagree over the exact legal inquiry that should determine whether a sexual tort falls within the scope of respondeat superior liability. They can be forgiven for doing so. The leading cases suggest what are, in effect, three different rules, and they do not explain how the different inquiries fit together. First, there is what the Court will call the “causal nexus” rule. Under that rule, one must ask whether a sexual assault was “engendered by” or an “outgrowth” of the employment. Id., 48 Cal.Rptr.2d 510, 907 P.2d at 363 (quoting Carr, 171 P.2d at 8). Lisa M. says that, for a sexual tort, this inquiry requires one to ask whether the tort's “motivating emotions were fairly attributable to work-related events or conditions.” Id., 48 Cal.Rptr.2d 510, 907 P.2d at 364. The necessary causal nexus is something more than “but for” causation: “[t]hat the employment brought tortfeasor and victim together in time and place is not enough.” Id., 48 Cal.Rptr.2d 510, 907 P.2d at 362.

Second, there is what the Court will call the “foreseeability” rule. It provides that respondeat superior liability applies where “ ‘[a] risk arises out of the employment when in the context of the particular enterprise an employee's conduct is not so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer's business.’ ” Mary M., 285 Cal.Rptr. 99, 814 P.2d at 1344 (quoting Perez v. Van Groningen & Sons, Inc., 41 Cal.3d 962, 227 Cal.Rptr. 106, 719 P.2d 676, 678 (1986)) (cleaned up); accord Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 362. In other words, respondeat superior liability applies “to the types of injuries that are ‘as a practical matter are sure to occur in the conduct of the employer's enterprise,’ ” or where the employment is “such as predictably to create the risk employees will commit intentional torts of the type for which liability is sought.” Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 362 (quoting Hinman, 88 Cal.Rptr. 188, 471 P.2d at 990).

Finally, there are the three policy factors—preventing recurrence, assuring compensation, and equitably spreading losses—described above. See Mary M., 285 Cal.Rptr. 99, 814 P.2d at 1343. One might think that these are not an independent “test,” and that instead they merely provide the underlying rationale for the “foreseeability” and “causal nexus” rules—a way to check one's work to confirm that those rules have been correctly applied. That is arguably how Lisa M. uses them. See 48 Cal.Rptr.2d 510, 907 P.2d at 366. But other cases do the opposite. They treat the policy factors as an independent part of the “scope of employment” analysis, and perhaps as actually determinative of it. In John R., for example, the Supreme Court of California held that the lower court had erred precisely because it “look[ed] mainly to the factual similarities” with prior cases instead of “consider[ing] whether the underlying justifications for the respondeat superior doctrine would be served by imposing vicarious liability here.” See John R., 256 Cal.Rptr. 766, 769 P.2d at 956. Whereas past cases might have dictated that a sexual tort was within the scope of employment, the John R. court preferred to ask whether the “underlying rationale for the respondeat superior doctrine” supported the imposition of vicarious liability. See id., 256 Cal.Rptr. 766, 769 P.2d at 955 (“But although the facts of this case can be made to fit a version of the respondeat superior doctrine, we are unpersuaded that they should be or that the doctrine is appropriately invoked here. We draw our decision not from the various factual scenarios in which vicarious liability has or has not been imposed on employers for the torts of their employees, but instead from the underlying rationale for the respondeat superior doctrine.”). Mary M. reasoned in a similar way. It quoted the foreseeability rule from Perez, but then applied that rule principally by looking to whether “the three policy objectives underlying respondeat superior would be achieved by applying the doctrine when a police officer on duty misuses his official authority and commits an act of rape.” Mary M., 285 Cal.Rptr. 99, 814 P.2d at 1347. Unlike in John R., the court in Mary M. decided that the policy rationale would be furthered by applying the doctrine to those facts, so it held that the defendant should be liable for its employee's sexual assault. See id.

In short, the leading cases leave it unclear how the causal nexus, foreseeability, or policy rules fit together—much less what should be done if they point in different directions. That ambiguity perhaps explains another case about which the parties disagree: Xue Lu v. Powell, 621 F.3d 944 (9th Cir. 2010). There, the Ninth Circuit offered its own synthesis of the law on respondeat superior liability for sexual torts in California. The case arose under the Federal Tort Claims Act, and it involved an asylum officer who had “abused his powers” to molest, and to solicit bribes from, asylum applicants he had met on the job and subsequently visited at home. Xue Lu, 621 F.3d at 949. The court surveyed most of the same cases as discussed here and described the “rule in California ․ as a form of enterprise liability,” emphasizing the issue of the “foreseeability of the employee's conduct.” Id. at 948; see also id. at 948–49. It held that the asylum officers’ torts were within the scope of his employment, emphasizing the policy factors and the fact that the conduct was “incidental to the asylum system.” Id. at 949. Uber points out, correctly, that Xue Lu was later criticized in an intermediate appellate case. See Z.V. v. Cnty. of Riverside, 238 Cal.App.4th 889, 189 Cal. Rptr. 3d 570, 577–81 (4th Dist. 2015) (discussing Xue Lu at length and concluding that the case “is not accurate either as a statement of California law or as an application of it.”). Some of Z.V.’s criticisms are persuasive: it is hard to argue that Xue Lu does not distort the “causal nexus” rule from Lisa M., and—despite deriving its legal standard from cases concerning assault and sexual assault—it ultimately relies on a factual comparison to a case about a real estate broker's fraud. But others are off base. For example, Z.V. offers a lengthy critique of Xue Lu’s reliance on the three policy factors, each of which it characterizes as “unhelpful in ascertaining respondeat superior liability” for various reasons. See id. at 580–81. But if that is the case, then Z.V.’s quarrel is not with the Ninth Circuit; it is with the Supreme Court of California, which treated those three policy questions as outcome-determinative in John R. and Mary M. Z.V. also ignores the ambiguity about the correct legal test that emerges from any fair reading of the cases. One cannot assert that the causal nexus rule is the beginning and end of the inquiry without ignoring two-thirds of Lisa M. and most of Mary M., John R., Farmers, and various intermediate appellate cases that focus on foreseeability.

If Lisa M. had wanted to make clear that the foreseeability and policy tests were subordinate to, or being replaced by, the causal nexus inquiry, it could have done so. It did not. Instead, it reaffirmed the “alternative focus” of foreseeability as a means of evaluating respondeat superior liability and the basic policy factors undergirding the case law.

b. Analysis

Applying the totality of this authority, the Court concludes that—at least at this stage—whether the alleged assaults were within the scope of the drivers’ employment presents a question of fact. See Doe v. Uber Techs., Inc., 184 F. Supp. 3d 774, 784 (N.D. Cal. 2016); but see Doe v. Uber Techs., Inc., No. 19-CV-03310-JSC, 2019 WL 6251189, at *3–*5 (N.D. Cal. Nov. 22, 2019). First, the foreseeability rule suggests that respondeat superior liability would be appropriate. It is certainly plausible, given the allegations about Uber's business model, that sexual assault is a risk “inherent in or created by the enterprise” that Uber has undertaken. John R., 256 Cal.Rptr. 766, 769 P.2d at 964 (concurrence). In reaching this conclusion, the Court emphasizes that foreseeability for purposes of respondeat superior is not the same as foreseeability for purposes of negligence. See Rodgers v. Kemper Constr. Co., 50 Cal.App.3d 608, 124 Cal. Rptr. 143, 148–49 (4th Dist. 1975). In this context, what matters is not “statistical frequency,” but the “relationship between the nature of the work involved and the type of tort committed.” Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 364. According to the MC, Uber's business model depends on pairing people who have cars with people who need rides. It also depends on making it very easy to become a driver—if it did not, Uber would be little different than a taxi service or a black car company. MC ¶¶ 137–41. The complaint alleges, plausibly, that a known risk of getting into a car with a stranger—for example, by hitchhiking—is the risk of sexual assault. Id. ¶ 1. It further alleges that Uber intentionally marketed its rides to populations who might be especially likely to be targeted for assault, including intoxicated people in need of rides home and women traveling alone. Id. ¶¶ 228–239. And Uber emphasized safety in its marketing over the years—including, for a time, stating that its drivers underwent “gold standard” or “industry leading” background checks. Id. ¶ 207. For present purposes, these allegations are relevant not because they show that Uber failed to exercise ordinary care or misleadingly marketed its services. They are relevant solely because they lend plausibility to Plaintiffs’ claim that sexual assaults are among those losses that “foreseeably result from the conduct of [Uber's] enterprise,” and that the tasks assigned to Uber's drivers—picking up strangers, often alone, sometimes intoxicated, sometimes late at night, in vehicles under the drivers’ sole control—are “such as predictability to create the risk” that some drivers would sexually assault the passengers. Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 362. No imaginative leaps are required to grasp why this is so.⁵

At least at this stage, the policy factors also point in favor of liability. Indeed, the policy factors distinguish this case from many of the cases where the California courts have declined to impose respondeat superior liability. Most of those cases concerned public employers or enterprises like hospitals, and the nature of those enterprises led courts to hesitate to impose liability where doing so might divert scarce resources from functions like providing healthcare or educating children, or where it might require the imposition of preventative measures that substantially interfered with those social goods. See Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 366–67; see also John R., 256 Cal.Rptr. 766, 769 P.2d at 956. The same concerns are not present here—at least, the Court cannot assume that they exist if it reads the pleadings in the light most favorable to Plaintiffs. Here, it is plausible that imposing liability on Uber “may prevent recurrence of the tortious conduct, because it creates a strong incentive for vigilance by those in a position to guard substantially against the evil to be prevented.” Mary M., 285 Cal.Rptr. 99, 814 P.2d at 1347 (quotation marks omitted). Uber might, for example, impose more stringent screening requirements for drivers, or it might add more robust monitoring or safety measures in vehicles.

It is also plausible, for obvious reasons, that imposing liability on Uber would increase the likelihood of compensation for victims of assault. In Lisa M., the court framed this inquiry as hinging on the availability of insurance “for sexual torts of employees,” and whether this would “drastically increase the insurance costs ․ of nonprofit providers such as Hospital.” Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 366. Here, a large, global firm like Uber could plausibly have access to custom insurance products that would cover claims like these, even if such products do not already exist (though of course the evidence may later show otherwise).

The final policy factor also weighs in favor of imposing liability. It concerns “the propriety of spreading the risk of losses among the beneficiaries of the enterprise upon which liability would be imposed,” which is “another way of asking whether the employee's conduct was ‘so unusual or startling that it would seem unfair to include the loss resulting from it among other costs of the employer's business.’ ” Id., 48 Cal.Rptr.2d 510, 907 P.2d at 367 (quoting Rodgers v. Kemper Constr. Co., 50 Cal.App.3d 608, 124 Cal. Rptr. 143, 149 (4th Dist. 1975)). For the reasons discussed above, the risk of sexual assault by rideshare drivers, given the nature of Uber's enterprise, is not “unusual or startling.” As for fairness, the Court cannot say that, as a matter of law, it would be unfair for Uber to have to count driver sexual assaults among the “costs” of its business. The MC describes a business model that depended precisely on making it extremely easy to become an Uber driver—something which has entailed quick, low-friction onboarding processes and few safety measures. See MC ¶¶ 137–41; see also id. ¶¶ 142–86. Again, this is not to confuse the negligence analysis with the respondeat superior analysis—it is just to get a grip on what Uber's “enterprise” actually is and what makes it distinct from, say, a taxi service or a school bus operator. Where an enterprise creates the risk that some drivers will sexually assault their passengers—where such assaults are “as a practical matter, sure to occur” because of what the enterprise is—the Court cannot say that it is unfair as a matter of law for Uber to have to internalize the costs of such assaults rather than shifting the risks to the passenger.

Finally, there is the causal nexus rule. This is the least favorable to Plaintiffs’ position, and Uber sets up its defense primarily on this point (and, more generally, on Lisa M.). In Lisa M., the plaintiff argued that her assault was engendered by work-related events or conditions because she was placed in a vulnerable position under the control of the ultrasound technician in an isolated environment. Lisa M. rejected the notion that the plaintiffs’ vulnerability alone could be a basis for imposing respondeat superior liability. See id., 48 Cal.Rptr.2d 510, 907 P.2d at 364 (“Here ․ a technician simply took advantage of solitude with a naive patient to commit an assault for reasons unrelated to his work. Tripoli's job was to perform a diagnostic examination and record the results. The task provided no occasion for a work-related dispute or any other work-related emotional involvement with the patient. The technician's decision to engage in conscious exploitation of the patient did not arise out of the performance of the examination, although the circumstances of the examination made it possible.”). Instead, the court emphasized that a causal nexus with a sexual assault would exist where the employee's “motivating emotions were fairly attributable to work-related events or conditions.” Id. Here, some of Plaintiffs’ arguments have concerned the vulnerability of riders, but Lisa M. forecloses this route. It is nevertheless possible to distinguish Lisa M. on the facts. The “conditions” of one's employment in a hospital are quite different than the conditions of one's employment as an Uber driver, not least because the conditions of employment by Uber include isolation, relative anonymity, extremely low barriers to entry, and an almost total lack of supervision. Being behind a closed door in a busy hospital is different.

It has to be acknowledged that if one takes literally the emphasis on the assailants’ “motivating emotions,” the causal nexus inquiry would likely counsel against imposing respondeat superior liability. Even if this is so, the Court would still allow the respondeat superior theory to proceed. The foreseeability and policy factors point towards liability, and where the result is not clear, the Supreme Court of California has suggested that the policy factors should ultimately guide the Court's decision. See John R., 256 Cal.Rptr. 766, 769 P.2d at 956; Mary M., 285 Cal.Rptr. 99, 814 P.2d at 1347. At any rate, it is impossible to read the full line of California cases on sexual torts and respondeat superior and to conclude that, in the final analysis, liability depends solely on some kind of psychological inquiry.⁶ Thus, for the reasons discussed above, the allegations are sufficient to create an issue of fact as to whether the alleged assaults were within the scope of employment under California law. Accord Doe v. Uber Techs., Inc., 184 F. Supp. 3d 774, 784 (N.D. Cal. 2016); but see Doe v. Uber Techs., Inc., No. 19-CV-03310-JSC, 2019 WL 6251189, at *3–*5 (N.D. Cal. Nov. 22, 2019).⁷

3. Apparent Agency

Plaintiffs’ second theory of vicarious liability is based on the doctrine of “apparent agency,” also called “ostensible agency.” Plaintiffs assert this theory in both California and Texas. Generally speaking, apparent agency “is based on the notion of estoppel, that is, a representation by the principal causing justifiable reliance and resulting harm.” Baptist Mem'l Hosp. Sys. v. Sampson, 969 S.W.2d 945, 948 (Tex. 1998). Here, Plaintiffs’ theory is that “because Uber represented that Uber drivers were its agents and because drivers with bad intentions used that apparent authority to get the plaintiffs alone with them in their cars, Uber is estoppel from maintaining now that passenger safety was the responsibility of the drivers, not Uber.” Omnibus Opp'n at 56–57 (citing MC ¶¶ 427–35). Put another way, Plaintiffs say that they reasonably relied on the appearance that their drivers were employed by Uber, a reputable company that presumably exercised a degree of control over its employees sufficient to deter them from assaulting passengers. As a result of that reliance, they got in the car with the drivers and were assaulted.

Uber does not dispute, for purposes of this motion, that Plaintiffs adequately allege facts showing that its drivers are its apparent agents. Omnibus Reply at 9 n.5. What the parties do dispute is the scope of vicarious liability created by an apparent agency relationship. Uber's view is that the scope of an apparent principal's liability can be at least no broader than the liability of an actual principal or employee. Accordingly, it argues that any vicarious liability is limited to torts committed within the scope of the apparent agency, and that, in determining what falls within that scope, the usual scope of employment rules apply.

Plaintiffs’ view, however, is that the scope of the vicarious tort liability of an apparent agent is unrelated to the scope of such liability for an employee. Instead of looking to the scope of the agency relationship, Plaintiffs argue that “the focus is on the connection between the tort and injury, on the one hand, and the reasonable reliance on the principal's representation, on the other hand.” Omnibus Opp'n at 58. In other words, according to Plaintiffs, “the principal is liable for injury that results from reliance on a false appearance the defendant wrongfully created,” and it is irrelevant whether the tort was within “scope” of the ostensibly agency. Id. at 59.

Apparent agency is a relatively well-established doctrine for imposing some kinds of tort liability on a principal who carries out his business using independent contractors, but who nevertheless creates (or knowingly permits) the appearance that the contractor is his employee. The doctrine is most common in the context of medical malpractice actions, and it has been recognized (in California, Texas, and elsewhere) as a viable theory for holding hospitals liable for the negligence of doctors who are held out as employees but are in fact independent contractors. See, e.g., Baptist Mem'l Hosp., 969 S.W.2d at 946. But Plaintiffs have not identified any cases where an agent was held liable for a tort under an apparent agency theory that would not have given rise to vicarious liability even if the tortfeasor had been an actual agent. Meanwhile, both the Restatement (Second) of Agency and the Restatement (Third) of Agency say that the scope of liability under this theory should be limited by the ordinary scope-of-employment rules. See Restatement (Second) of Agency § 267, cmt. b (1958); Restatement (Third) of Agency § 7.08, n. b (2006). In the only cited case that has directly confronted the question, Judge Corley answered in the negative. See Doe v. Uber Technologies, Inc., 2020 WL 2097599, at *1 (N.D.Cal. May 1, 2020). That seems to be the best answer under the laws of both California and Texas: the scope of liability under an apparent agency theory is limited by the usual scope-of-employment rules.

It is true that the legal tests for “scope of employment” liability and “apparent agency” liability are different, and that apparent agency is based on estoppel. But, as Judge Corley held in Doe v. Uber, “[w]hile it is true that the ‘[l]iability of the principal for the acts of an ostensible agent rests on the doctrine of ‘estoppel,’ it does not follow that an employer could be liable for acts of an ostensible agent, but not the same acts committed by an employee or agent.” 2020 WL 2097599, at *1 (quoting Kaplan v. Coldwell Banker Residential Affiliates, Inc., 59 Cal. App. 4th 741, 747, 69 Cal.Rptr.2d 640 (1997)). To see why, it helps to note that while the cases do not expressly require an inquiry into the scope of apparent agency, they do require that the plaintiff have a reasonable belief in the agent's authority, and that this belief must cause the injury. See, e.g., Kaplan, 59 Cal. App. 4th at 747, 69 Cal.Rptr.2d 640 (“[T]here are three requirements necessary before recovery may be had against a principal for the act of an ostensible agent. [1] The person dealing with the agent must do so with belief in the agent's authority and this belief must be a reasonable one; [2] such belief must be generated by some act or neglect of the principal sought to be charged; [3] and the third person in relying on the agent's apparent authority must not be guilty of negligence.”) (internal quotation marks omitted) (emphasis added).⁸

It is indisputable that “ ‘apparent authority must be based on the acts of the principal’ and ‘is limited to the scope of responsibility that is apparently authorized.’ ” Gaines v. Kelly, 235 S.W.3d 179, 184 (Tex. 2007) (quoting First Valley Bank of Los Fresnos v. Martin, 144 S.W.3d 466, 471 (Tex. 2004)); accord Blanton v. Womancare, Inc., 38 Cal.3d 396, 212 Cal.Rptr. 151, 696 P.2d 645, 651 (1985) (“[A]pparent authority is created, and its scope defined, by the acts of the principal in placing the agent in such a position that he appears to have the authority which he claims or exercises.”). This implies a limit to the estoppel idea on which Plaintiffs hang their hats. For example, where the scope of an agent's apparent authority was limited to delivering papers rather than cutting deals, a plaintiff who cuts a fraudulent deal with the agent cannot hold the principal vicariously liable for the fraud. The transaction, and therefore the fraud, was beyond the scope of the apparent authority. See Gaines, 235 S.W.3d at 185.

If the type of apparent authority conferred on an apparent agent is that of an employee, the same limitation applies. An employee is a special type of agent, but employees’ authority is not unlimited, and no one could reasonably believe it to be so. See, e.g., Clark Equip. Co. v. Wheat, 92 Cal.App.3d 503, 154 Cal. Rptr. 874, 883 (1st Dist. 1979) (holding that vicarious liability could lie for the act of an apparent agent where “[f]rom the point of view of the third person the transaction seems regular on its face and the agent appears to be acting in the ordinary course of the business confided to him”) (emphasis added) (internal quotation marks removed). Consider a case like Baptist Mem'l Hospital, where a hospital was held vicariously liable for the medical malpractice of a doctor who it held out as its agent, because there could be no question that the doctor's act would be within the scope of his apparent employment. See 969 S.W.2d at 946. But if the plaintiff had accepted a ride home from the doctor after her operation, and the doctor crashed his car, then an obvious question about scope—or in other words, about the extent of apparent authority—would arise, even if the doctor seemed to be an employee of the hospital. One would have to ask: Did the hospital do anything to suggest that the scope of the doctor's authority including driving patients home? The same would be true of an Uber driver who, although an apparent agent, would probably not be acting within the scope of his apparent authority if he performed an appendectomy on a rider. In short, the idea of “scope” does not come into the inquiry in the same way as in the respondeat superior analysis, but limited scope is implicit in the idea of apparent authority.⁹

It follows that the scope of vicarious liability for the torts of an agent who apparently has some “business confided to him” can be no greater than the scope of such liability for an agent who actually has that “business confided to him.” Clark Equip. Co., 154 Cal. Rptr. at 883; accord Doe, 2020 WL 2097599, at *1; Restatement (Second) of Agency § 267, cmt. b (1958); Restatement (Third) of Agency § 7.08, n. b (2006).¹⁰ This does not doom Plaintiffs’ apparent agency theory in California, where Plaintiffs have plausibly pled that Uber could be vicariously liable for the driver's alleged torts. But it does doom that theory in Texas, where Plaintiffs concede that Uber could not be vicariously liable under a theory of respondeat superior. Accordingly, the “apparent agency” theory is dismissed as to the Texas plaintiffs.

4. Ratification

Plaintiffs’ next theory of vicarious liability is based on the agency law doctrine of ratification. Generally speaking, even if the assaults were outside of the drivers’ actual authority, Uber “ratified” the drivers’ acts when it “failed to respond adequately to reports of sexual misconduct, failed to adequately investigate reports of sexual misconduct, failed to report sexual misconduct to law enforcement, and reinstated drivers after complaints of sexual misconduct.” MC ¶ 438.

Uber disputes Plaintiffs’ view of the applicable legal standards in California and Texas as discussed below. The Court first discusses those issues, and then proceeds to Uber's arguments about factual deficiencies in Plaintiffs’ pleading.

a. California Law on Ratification

In California, “[a] purported agent's act may be adopted expressly, or it may be adopted by implication based on conduct of the purported principal from which an intention to consent or to adopt the act may be fairly inferred ․” Dickinson v. Cosby, 37 Cal.App.5th 1138, 250 Cal. Rptr. 3d 350, 366 (2d Dist. 2019) (internal quotation marks omitted). Conduct supporting that intention includes “conduct which is inconsistent with any reasonable intention on his part, other than ․ approving and adopting it.” Id. (internal quotation marks removed). In California, ratification is a well-established doctrine for assigning vicarious liability to a principal for torts like assault or battery, even where those torts are outside of the scope of the agency relationship. See Thomas v. Regents of U. of California, 97 Cal.App.5th 587, 315 Cal. Rptr. 3d 623, 648 (1st Dist. 2023) (“The theory of ratification is generally applied where an employer fails to investigate or respond to charges that an employee committed an intentional tort, such as assault or battery.”) (quoting C.R. v. Tenet Healthcare Corp., 169 Cal.App.4th 1094, 87 Cal. Rptr. 3d 424, 437 (2d Dist. 2009)). “An employer's ‘failure to investigate or respond to charges that an employee has committed an intentional tort’ or ‘failure to discharge the employee may be evidence of ratification’ by the employer.” Id. at 649 (quoting Samantha B. v. Aurora Vista Del Mar, LLC, 77 Cal.App.5th 85, 292 Cal. Rptr. 3d 324, 342 (2d Dist. 2022)); see also C.R., 87 Cal. Rptr. 3d at 437 (“[R]atification may occur when an employer learns of misconduct and fails to discharge an agent or employee.”). “Whether an employer has ratified an employee's conduct is generally a factual question.” Thomas, 315 Cal. Rptr. 3d at 648.

b. Texas Law on Ratification

“Under Texas law, an employer may be vicariously liable for the intentional tort of its employee under the doctrine of respondeat superior or directly liable under the theory of ratification.” Skidmore v. Precision Printing and Pkg., Inc., 188 F.3d 606, 614 (5th Cir. 1999). In tort cases, “ratification may occur when the employer or its vice-principal confirms, adopts, or fails to repudiate the acts of its employee.” Prunty v. Arkansas Freightways, Inc., 16 F.3d 649, 653 (5th Cir. 1994). The Fifth Circuit has read Texas law on this point as follows: “When the company 1) knows about the employee's acts, 2) recognizes that the employee's acts will continue if he is retained, 3) does nothing to prevent the ongoing tortious acts, and 4) chooses to retain the employee, the company ratifies the tortious acts and may be held liable for exemplary damages.” Id. at 653–54; see also Verhelst v. Michael D's Rest. San Antonio, Inc., 154 F. Supp. 2d 959, 966 (W.D. Tex. 2001) (triable issue of fact on ratification of sexual assault where the principal was fully informed, took affirmative steps to ensure that the plaintiff remained in peril, and stated “he did not want to hear anything else about it”). By themselves, “neither the mere retention of an employee or denial of liability establish ratification.” Id. Instead, the principal “must perform a voluntary, intentional act which is inconsistent with an intention of avoiding” the prior act. Old Rep. Ins. Co., Inc. v. Fuller, 919 S.W.2d 726, 728 n.1 (Tex. App. 1996).

c. Pleading Issues

Aside from disputing the correct legal standard, Uber argues that, to support a ratification theory, Plaintiffs have to plead facts showing that Uber ratified the misconduct of specific agents in connection with specific incidents—for example, by reinstating or failing to fire a particular driver who had been reported for misconduct.

The Court agrees that the facts alleged are insufficient to support a theory of ratification. The allegations about “Uber's knowledge of an epidemic of sexual assault in its transportation system, and its pervasive pattern of gross indifference to that misconduct,” if true, certainly mean that it is possible that Uber “would have known about each of the alleged assaults and would have applied its standard policies in response.” Omnibus Opp'n at 49–50 (citing MC ¶¶ 240–316). But just by themselves, the allegations do not support a plausible inference that Uber knew about any given incident, nor do they permit inferences about what actions Uber took with respect to particular drivers. (Indeed, some of the allegations cut against the ratification theory, since they tend to suggest that Uber deliberately avoided learning about particular incidents of sexual assault or that victims often did not report them for any number of reasons. Those allegations may help with Plaintiffs’ other theories, but they suggest that whether Uber learned about and ratified a given incident can't be decided without reference to individual allegations.) Plaintiffs are correct that much of the relevant information is exclusively in Uber's possession, but not all of it is. For example, it could probably be inferred that Uber knew about a given incident if it was reported to Uber or to law enforcement, but Plaintiffs have not pointed to any such allegations in the short-form complaints. Some plaintiffs may receive information that will allow them to plead this theory under California or Texas law—the information in the Defense Fact Sheets may help—and they could seek to amend their complaints at the appropriate time. But for now, the ratification theory has not been adequately pled, and is dismissed on that basis.

5. Common Carriers’ Non-Delegable Duties

Finally, Plaintiffs argue that Uber's non-delegable duties as a common carrier create an independent basis for vicarious liability. Under Texas law, this road is closed to Plaintiffs, for the Texas legislature has expressly declared that “[t]ransportation network companies and drivers logged in to the company's digital network are not common carriers, contract carriers, or motor carriers.” Tex. Occ. Code Ann. § 2402.002.¹¹ Accordingly, Plaintiffs’ Claim E is dismissed under Texas law.

Uber does not dispute that it is a common carrier and explicitly notes in its Reply that it did not move to dismiss Plaintiffs’ Claim E under California law. Omnibus Reply at 73. As such, the Court need not analyze Plaintiffs’ Claim E for the purposes of the California Motion.

But Uber does contest Plaintiffs’ position that Uber's status as a common carrier gives rise to an independent basis for vicarious liability. There, Uber's arguments are unconvincing. Under traditional common law principles, common carriers have duties to safely transport passengers. See Restatement (Second) of Torts § 314A. In California, this duty is codified in Cal. Civ. Code § 2100, which provides that “[a] carrier of persons for reward must use the utmost care and diligence for their safe carriage, must provide everything necessary for that purpose, and must exercise to that end a reasonable degree of skill.” The Supreme Court of California has described common carriers’ duties as nondelegable. Eli v. Murphy, 39 Cal.2d 598, 248 P.2d 756, 757 (1952). Instead, the carrier's duty is only satisfied if its agent actually does exercise the requisite level of care. See Restatement (Second) of Agency § 214, cmt. a.

Plaintiffs (and some cases) characterize this as essentially a form of vicarious liability. See, e.g., Srithong v. Total Inv. Co., 23 Cal.App.4th 721, 28 Cal. Rptr. 2d 672, 675 (2d Dist. 1994) (“[T]he nondelegable duty rule is a form of vicarious liability because it is not based on the personal fault of the landowner who hired the independent contractor.”). Uber disputes that characterization, and it suggests instead that the doctrine is only a form of direct liability—one that would, presumably, create no liability on Uber's part beyond what it would otherwise have under Plaintiffs’ negligence theories. But that argument depends on a misunderstanding of non-delegable duties. “Vicarious liability means that the act or omission of one person is imputed by operation of law to another.” Srithong, 28 Cal. Rptr. 2d at 675 (cleaned up). It is axiomatic that, where a duty is nondelegable, one cannot avoid liability by claiming that discharge of the duty was somebody else's responsibility. See SeaBright Ins. Co. v. US Airways, Inc., 52 Cal.4th 590, 129 Cal.Rptr.3d 601, 258 P.3d 737, 743 (2011). The question is not whether the principal exercised sufficient care in selecting the agent or supervising its work; that would be a direct liability claim (and it would be premised on a different tort duty). The question is simply whether the duty was actually discharged. For example, “car owners cannot delegate their duty to ensure that their cars have working brakes,” which means that “an owner cannot avoid liability for an accident by arguing that the mechanic hired to inspect the brakes failed to discover the brake problem.” Id. (discussing Maloney v. Rath, 69 Cal.2d 442, 71 Cal.Rptr. 897, 445 P.2d 513 (1968)). The relevant duty is not the duty to exercise reasonable care in selecting a competent mechanic; it is the duty to have working brakes. In that example, liability is “vicarious” in the sense that the mechanic's negligence is effectively imputed to the car owner, who may have acted perfectly reasonably in entrusting the brakes to the mechanic's care.

So here, Uber, as a common carrier, must “use the utmost care and diligence for [passengers’] safe carriage, must provide everything necessary for that purpose, and must exercise to that end a reasonable degree of skill.” Cal. Civ. Code § 2100. If instead of exercising the utmost care and diligence for passenger's safe carriage, a driver assaults the passenger, then Uber's non-delegable duty has been breached. That conclusion follows from Berger v. S. Pac. Co., 144 Cal.App.2d 1, 300 P.2d 170 (1st Dist. 1956). There, a passenger in a Pullman sleeping car alleged that she was sexually assaulted by the car's porter. The court held that the jury had been properly instructed that if it found that the assault had been committed and that the porter was an agent of the defendant, then the defendant was liable. It quoted with approval the following statement of the law:

A carrier is liable for acts of assault and battery upon the part of its employees resulting in injury to those it has agreed to transport upon its facilities. This liability extends not only to cases where the assault was in the line of the employee's duty, but also to those instances where the act was merely that of an individual and entirely disconnected with the performance of the agent's duties, as where the conductor of a train kisses a female passenger against her will. * * * The liability of a common carrier for an assault by one if its employees on a passenger is not dependent on the question as to whether the employee was acting within the scope of his authority or in the line of his duty, but is based upon its broad duty as a common carrier to protect its passengers from assault.

Id. at 173.

Uber argues that Berger is outdated, wrongly decided, and that recognizing this rule would essentially override the scope-of-employment rule set out in Lisa M. But these contentions are unpersuasive. First, Uber relies on subsequent cases that concern assaults on passengers not by the agents of the common carrier, but by third parties (generally, fellow passengers). See Lopez v. S. Cal. Rapid Transit Dist., 40 Cal. 3d 780, 785, 221 Cal.Rptr. 840, 710 P.2d 907 (1985). It is in that connection in which the Supreme Court of California said that common carriers “are not ․ insurers of their passengers’ safety.” Id. That principle makes a good deal of sense; a heightened duty of care does not mean that a common carrier is strictly liable for anything that happens to its passengers. See Maloney v. Rath, 69 Cal.2d 442, 71 Cal.Rptr. 897, 445 P.2d 513, 515 (1968). In other words, a common carrier's duty of care is not breached every time a passenger gets hurt in transit, no matter how it happens. But that is not the basis for the Plaintiffs’ theory in this case. Instead, Plaintiffs’ theory is that, when a driver assaults a passenger, there is no question that the driver has failed to discharge the common carriers’ duties: instead of safely transporting the passenger, the common carrier's agent intentionally assaulted her. That follows from Berger, and properly understood, it is in no tension with Lopez.

Second, it would be one thing to suggest that Berger was outdated if there was California authority to the contrary, but there is not. Instead, California continues to hold common carriers to a heightened standard of care by statute. See Cal. Civ. Code § 2100. Berger is directly on point, has never been overruled, and does not ever seem to have been criticized by California courts. See Doe v. Uber Techs., Inc., 184 F. Supp. 3d 774, 787 (N.D. Cal. 2016) (rejecting effort to distinguish Berger where Uber “relied largely on cases addressing negligence or passenger-on-passenger assault, and on a case decided by the New York Court of Appeals,” and noting that “[t]hese cases simply are not analogous to the present facts, where plaintiffs have alleged that an employee intentionally sexually assaulted a passenger”).

Finally, it is not true that recognizing vicarious liability under this theory would somehow be inconsistent with Lisa M., even if Uber were correct that Lisa M. precluded vicarious liability for its drivers’ sexual torts. Lisa M. simply addresses one possible theory of vicarious liability—respondeat superior or “scope of employment” liability. It does not address other forms of vicarious liability, including such liability where it might arise from nondelegable duties. See California Assn. of Health Facilities v. Dep't of Health Services, 16 Cal.4th 284, 65 Cal.Rptr.2d 872, 940 P.2d 323, 336 (1997) (“Finally, CAHF argues that the recent holding in [Lisa M.] is an indication of an intent to limit or repudiate the doctrine of vicarious liability and nondelegable duties. This is not the case․ Nothing in Lisa M. indicates a retreat from the rule of nondelegable duties for licensees.”). Certainly, it follows from Berger that the scope of a common carrier's liability for its agents’ intentional torts could be broader than that of, say, a hospital. But that is just a reflection of the fact that common carriers have non-delegable duties, while most other employers do not. As such, Plaintiffs’ theory that a common carrier's non-delegable duties can form an independent basis for vicarious liability can proceed under California law. See Doe v. Uber Techs., Inc., 184 F. Supp. 3d 774, 787 (N.D. Cal. 2016).

C. “Other Non-Delegable Duties” Claim

Plaintiffs concede that Claim F, which “is premised on Uber's non-delegable duties that do not arise from its status as a common carrier,” must be dismissed under California and Texas law. Hearing Tr. (dkt. 615) at 44:4–6.

D. Fraud-Related Claims

The fraud-related claims are premised on allegations that Uber made both fraudulent misrepresentations and fraudulent omissions about the safety of its rides, and that “[a]s a result of relying on Uber's misrepresentations and omissions, Plaintiffs were assaulted or harassed by Uber drivers[.]” MC ¶¶ 368–80. The parties raise a variety of arguments about the viability of the fraud claims. At this stage, the Court need only reach one: the lack of particular allegations about who saw and relied on which representations and when they saw them.

To satisfy Rule 9(b), allegations of fraud “must be accompanied by ‘the who, what, when, where, and how of the misconduct charged.’ ” Kearns v. Ford Motor Co., 567 F.3d 1120, 1125 (9th Cir. 2009). “ ‘[U]nadorned references to [an] advertising campaign and marketing materials’ do not meet these requirements.” In re Volkswagen “Clean Diesel” Mktg., Sales Practices, and Products Liab. Litig., 349 F. Supp. 3d 881, 914 (N.D. Cal. 2018) (quoting Yastrab v. Apple Inc., 173 F. Supp. 3d 972, 978 (N.D. Cal. 2016)). Instead, pleading a fraud-based claim requires “linking allegations” that connect plaintiffs to allegedly false or misleading marketing. See id. at 916. In other words, a plaintiff must allege that she “saw these advertisements and relied on the promises contained within them.” Id.; see also Kearns, 567 F.3d at 1126 (affirming dismissal of fraud-based claims under Rule 9(b) in part because the plaintiff failed to “specify when he was exposed to [marketing and advertising materials] or which ones he found material”).

While the MC contains numerous specific examples of Uber's advertising that it says are fraudulent, Plaintiffs do not dispute that they have not pled any plaintiff-specific allegations about whether a given plaintiff saw certain alleged misrepresentations, how they relied on them, and so forth. For the time being, Plaintiffs essentially ask the Court to examine the totality of Uber's messaging about safety and determine that, as a whole, it was fraudulent. See Omnibus Opp'n at 68 (“Plaintiffs’ fraud claims are not about isolated sentences, but about a decade-long torrent of representations both general and specific, clear and vague, express and implied, and all in the context of a business whose very existence depends on passengers believing in and relying on those representations.”).

But it is not possible to assess any of the fraud claims without knowing what representations a particular person saw and relied on. See, e.g., Graham v. Bank of Am., N.A., 226 Cal.App.4th 594, 172 Cal. Rptr. 3d 218, 228 (4th Dist. 2014); In Interest of C. M. V., 479 S.W.3d 352, 361 (Tex. App.--El Paso 2015). Here, Plaintiffs say that the fraud claims depend on “the duration (more than a decade) and pervasiveness (high) of Uber's misrepresentations,” and that “more generalized representations came on the heels of, and reinforced” certain more egregious misrepresentations that were made during an earlier period. Opp'n at 68. This suggests that the problem may actually be deeper than the absence of allegations about reliance. If the alleged statements are fraudulent because of their contextualization with other statements over a period of time, then it obviously matters what representations a particular plaintiff saw and when she saw them. (Presumably no actual person saw and relied on the entirety of Uber's messaging about safety over the course of ten years.) For example, someone who was assaulted in 2023 may have only been aware of the things Uber was saying about safety starting in 2022, and the “context” of the representations on which someone assaulted in 2017 could have relied obviously would not have included the entire “decade-long torrent” of representations. Plaintiffs defend the claim by saying that “[t]he Master Complaint pleads facts that make it plausible that these claims, as a category, are live and defensible, and applicable to large segments of the cases pending (or that will be pending) in the MDL.” Omnibus Opp'n at 86. But neither Rule 8, Rule 9, nor any authority cited by Plaintiffs contemplates anything like plausibly pleading a “category” of claims. As discussed above, this is no less true in MDL proceedings. See In re Korean Air Lines Co., 642 F.3d at 700.

The problem extends to the claims based on fraudulent omissions. Those claims turn in significant part on rules recognizing duties to disclose that arise from telling half-truths. See Bombardier Aerospace Corp. v. SPEP Aircraft Holdings, LLC, 572 S.W.3d 213, 219–20 (Tex. 2019) (duty to disclose where “the defendant: (1) discovered new information that made its earlier representation untrue or misleading; (2) made a partial disclosure that created a false impression; or (3) voluntarily disclosed some information, creating a duty to disclose the whole truth”); Kulp v. Munchkin, Inc., 678 F. Supp. 3d 1158, 1169 (C.D. Cal. 2023) (California duty to disclose where “the defendant makes partial representations but also suppresses some material facts”). Assessing whether Uber owed a given plaintiff a duty to disclose based on allegedly misleading partial representations depends on what the relevant partial representations were—which, again, depends on what a given plaintiff actually saw and relied upon.

Accordingly, this claim is dismissed with leave to amend.

E. Negligent Infliction of Emotional Distress

Uber argues that Plaintiffs’ separate claim for negligent infliction of emotional distress (NIED) should be dismissed because neither California nor Texas recognizes NIED as an independent cause of action. In their opposition, Plaintiffs withdraw their Texas NIED claim. Opp'n at 3. And in California, Uber is correct that “[t]he negligent causing of emotional distress is not an independent tort, but the tort of negligence.” Burgess v. Super. Ct., 2 Cal.4th 1064, 9 Cal.Rptr.2d 615, 831 P.2d 1197, 1200 (1992).

To be clear, Uber has not moved to dismiss the plaintiffs’ core negligence claims, nor does it argue that Plaintiffs haven't pled facts that would entitle them to recover for emotional distress under the rubric of ordinary negligence. But insofar as Plaintiffs assert NIED as an independent cause of action under California law, that claim is dismissed.

F. Negligent Entrustment

Plaintiffs’ negligent entrustment theory is that Uber permits “drivers to use its chattel, the Uber product and trademarks,” including Uber's “trademarked decals and signage” that the drivers hang on their vehicles. Omnibus Opp'n at 90-91; see MC ¶¶ 78, 192-93, 13, 363(n). The idea is that drivers permitted to use these items can exploit “the imprimatur of a legitimate and responsible corporation,” which gives “passengers the necessary sense of safety and security to get into the car with an Uber driver.” MC. ¶ 13.

This claim is due to be dismissed under the laws of both states. First, in both states, Uber is correct that the negligent entrustment cases uniformly concern instrumentalities that are themselves at inherently dangerous in some way. This is clearest in Texas, where the courts have been parsimonious about expanding the tort beyond the context of vehicles, and where the tort is always defined in vehicle-specific terms. See 4Front Engineered Sols., Inc. v. Rosales, 505 S.W.3d 905 (Tex. 2016); Goodyear Tire & Rubber Co. v. Mayes, 236 S.W.3d 754, 758 (Tex. 2007); Dee v. Parish, 160 Tex. 171, 327 S.W.2d 449, 452 (1959). The farthest afield that the intermediate appellate courts have gone is in applying the tort to firearms, although not all of them have been willing to do so. See Kennedy v. Baird, 682 S.W.2d 377, 378 (Tex. App.--El Paso 1984) (noting that “[a] firearm is a deadly weapon per se” as justification for extending negligent entrustment tort to encompass them); Prather v. Brandt, 981 S.W.2d 801 (Tex. App.--Hous. [1st Dist.] 1998); but see Richardson v. Crawford, No. 10-11-00089-CV, 2011 WL 4837849, at *4 n.5 (Tex. App.--Waco Oct. 12, 2011) (explaining that the decision “should not be interpreted to mean that we recognize ‘negligent entrustment’ to include personal property, in general, or of a firearm, in particular, as a cause of action”). Given that legal background, it seems unlikely that the Texas Supreme Court would extend negligent entrustment liability for branding materials like decals and signs when intermediate appellate courts have hesitated even to apply the tort to guns.

The same is true in California, although the reasoning is slightly different. California courts have at least suggested that the separate tort of negligent entrustment applies only to “dangerous instrumentalit[ies].” See Jacoves v. United Merch. Corp., 9 Cal. App. 4th 88, 116, 11 Cal.Rptr.2d 468 (Cal. Ct. App. 1992) (“A person owes a duty of care not to provide a dangerous instrumentality to an individual whose use of the instrumentality the supplier knows, or has reason to know, will result in injury.”). And California courts have adopted the definition of the tort set out in Section 390 of the Restatement (Second) of Torts, which—although it does not explicitly limit the tort to inherently dangerous chattels—certainly does not contemplate a claim like plaintiffs’ involving decals and stickers. See id. at 115, 11 Cal.Rptr.2d 468 (“The California Supreme Court in Richards v. Stanley, [43 Cal.2d 60, 271 P.2d 23 (1954)], adopted the law set forth in Restatement Second of Torts section 390”); Restatement (Second) of Torts § 390, cmts. a–d (discussing illustrations that all concern cars, boats, or guns). Moreover, where California courts have accepted legal theories closer to Plaintiffs’ entrustment theory, they have done so under the framework of the general negligence statute, not the separate framework of negligent entrustment. Specifically, in Hacala v. Bird Rides, Inc., 90 Cal.App.5th 292, 306 Cal. Rptr. 3d 900 (2d Dist. 2023), review denied (June 21, 2023), the California Court of Appeal held that Bird, an electric scooter company, owed a duty “not to entrust its scooters to individuals who Bird knows or should know are likely to leave scooters in hazardous locations where they will pose an unreasonable risk of harm to others.” Id. at 919. It rooted that holding in the general duty codified in Cal. Civ. Code § 1714 “to use ordinary care in the management of [one]’s property. Hacala, 306 Cal. Rptr. 3d at 916; see also Cal. Civ. Code § 1714(a) (“Everyone is responsible, not only for the result of his or her willful acts, but also for an injury occasioned to another by his or her want of ordinary care or skill in the management of his or her property or person, except so far as the latter has, willfully or by want of ordinary care, brought the injury upon himself or herself.”). If the California Supreme Court would recognize a theory like Plaintiffs’, it would likely be under the rubric of this general duty of care, not under the separate rubric of negligent entrustment.

Insofar as Plaintiffs assert negligent entrustment under a separate cause of action in Texas and California, that claim is dismissed.¹²

G. Strict Products Liability

Plaintiffs allege that the Uber App is a defective product, and that its defects caused them to suffer their injuries. Plaintiffs acknowledge that “Uber's business model is built around hybrid product-services transactions,” and that these transactions involve “a product (the App) placed in the stream of commerce through which users obtain services (Uber rides).” Omnibus Opp'n at 93 (citing MC ¶¶ 51, 445, 449, 451-53, 477, 480, 491, 494). Plaintiffs nevertheless argue that defects in the Uber App can subject Uber to strict liability. Specifically, they allege that Uber “failed to design away the risks of sexual harassment, assault, kidnapping, or other sexual misconduct while using the Uber app,” for example by:

a. Failing to design an effective safety-focused reporting system to capture safety data so it could be acted on;

b. Failing to permit passengers to select drivers of the same gender;

c. Failing to enact and appropriately enforce a zero tolerance policy toward drivers with a history of inappropriate behavior;

d. Failing to appropriately monitor rides to detect known patterns of sexual assault, harassment, kidnapping or other forms of sexual assault, including but not limited to using existing GPS and predictive technology to monitor and respond to route deviations or drivers spending excessive time with passengers at the start or end of a route;

e. Failing to offer timely support during unsafe rides; and

f. Failing to add appropriate background checks before and after allowing drivers onto the app, including but not limited to biometric screening.

MC ¶¶ 472–473. Plaintiffs also allege that the app is defective under a failure to warn theory, because Uber “Uber failed to provide adequate warnings to users of its app by failing to warn them to the actual risk of sexual harassment, assault, kidnapping, or other sexual misconduct.” Id. ¶ 474.

1. Whether the Uber App Is a Product

The threshold question, which is hotly disputed, is whether the Uber app is a product in the first place. In her social media addiction MDL, Judge Gonzalez Rogers recently observed that “applicability of products liability torts to the digital world” presents “novel questions of law.” In re Social Media Adolescent Addiction/Pers. Inj. Prods. Liab. Litig., 702 F.Supp.3d 809, 838 (N.D. Cal. 2023). That is certainly true.

The parties agree that, in California and Texas, the Restatement (Third) of Torts: Prods. Liab. § 19 provides the applicable definition of a “product.” It provides:

(a) A product is tangible personal property distributed commercially for use or consumption. Other items, such as real property and electricity, are products when the context of their distribution and use is sufficiently analogous to the distribution and use of tangible personal property that it is appropriate to apply the rules stated in this Restatement.

(b) Services, even when provided commercially, are not products.

The Restatement takes no official position on whether software should be considered a product. The commentary, however, suggests that the reporters believed that it probably should be, at least where it is mass produced rather than custom made for a particular user. The Reporter's Notes state that “numerous commentators have discussed the issue and urged that software should be treated as a product,” and it suggests that courts considering the issue should take guidance from decisions that have held that mass-marketed software is a “good” under the Uniform Commercial Code. See id. § 19, Reporter's Notes, cmt. d. The Ninth Circuit once suggested in dicta “[c]omputer software that fails to yield the result for which it was designed” might appropriately be recognized as a defective product. Winter v. G.P. Putnam's Sons, 938 F.2d 1033 (9th Cir. 1991). Nevertheless, the question has not been definitively answered. Uber cites a handful of decisions (almost all unpublished cases from state trial courts) that have held that the Uber app is not a product. See Cal Mot. at 24 n.18. Plaintiffs point to a couple of other state trial court decisions that have gone the other way. See Omnibus Opp'n at 95–96.

At this stage, the Court is persuaded that the Uber app would be considered a product under the Restatement. It bears noting, however, that the Court finds Plaintiffs’ argument that the app is “tangible”—in the sense that downloading the app has tiny, but real, physical effects on the world ¹³ —to be too clever by half. “Tangible” does not really mean “having some effect on physical reality.” At any rate, there is no reason to strain language in this way. The better argument is that the “context of [the app's] distribution and use is sufficiently analogous to the distribution and use of tangible personal property that it is appropriate to apply the rules” of strict liability. Restatement (Third) of Torts: Prods. Liab. § 19. The app's frontend interface is designed entirely by Uber, distributed by Uber to app stores, and then—to access it—a user has to download the app onto their personal mobile device from the relevant app store. MC ¶¶ 446–48. The use of the app is analogous to the use of any tangible appliance one might bring home.¹⁴ If instead of a mobile phone application, Uber manufactured a mobile phone accessory—say, a little button that plugged into a smartphone and which, if pressed, would summon an Uber ride to the user's location—no one would argue that the accessory was not a product. Uber has not offered any arguments about the app's distribution and use that do not boil down to an outdated fetishization of physical objects. The law is not so sclerotic as that.¹⁵

This reasoning finds further support in the “defect-specific approach” used by Judge Gonzalez Rogers in analyzing social media apps. See In re Social Media Adolescent Addiction/Pers. Inj. Prod. Liab. Litig., 702 F.Supp.3d 809, 848–54 (N.D. Cal. 2023). There, Judge Gonzalez Rogers analogized the functionalities of the apps to tangible products, including comparing defects in parental controls and age verification systems in the apps to real-world products like physical parental locks on prescription medicines and software parental locks on televisions. The real-world parental locks are undeniably products, so their app-based analogues are too.

Here, Plaintiffs allege defects in Uber's app that have similar plausible analogues in tangible products, like Plaintiffs’ allegation that Uber failed to implement GPS-based safety alerts to flag route deviations. GPS-based alerting and tracking exists as part of many tangible products and can be used for similar purposes. For example, Apple AirTags alert a user if the device is left behind at an unauthorized location. The Uber app's alleged defect is the failure to use GPS to do just what a tangible product like an Apple AirTag does, and thus is also a product. As in In re Social Media Adolescent Addiction/Pers. Inj. Prod. Liab. Litig., those allegations suffice to establish at least some of the alleged defects as products. Thus, at the pleadings stage, Plaintiffs’ product liability claims will not be dismissed on this basis.¹⁶

2. Predominant Purpose

Uber argues that, even if the app is a product, “[c]ourts have not extended the doctrine of strict liability to transactions whose primary objective is obtaining services.” Ferrari v. Grand Canyon Dories, 32 Cal.App.4th 248, 38 Cal. Rptr. 2d 65, 71 (3d Dist. 1995). As Plaintiffs acknowledge, the point of downloading the Uber app is to get a ride from an Uber driver, and a ride is a service, not a product. Axiomatically, services are not products and do not give rise to strict liability. Restatement (Third) of Torts: Prods. Liab. § 19(b).

But the cases concerning what Uber calls the “predominant purpose test” simply do not mean what Uber would like them to. None of the cases Uber cites on this point concern the “predominant purpose” of the product itself; they do not say that a product used for procuring a service (or in connection with a transaction for services) is thereby excluded from products liability law. Instead, these cases concern who in the chain of distribution of a product can be liable under a product defect theory—that is, who “sells or otherwise distributes” a product. Restatement (Third) of Torts: Prods. Liab. § 20; see also Ferrari, 38 Cal. Rptr. 2d at 70 (discussing who “strict liability has been imposed against”); Hennigan v. White, 199 Cal.App.4th 395, 130 Cal. Rptr. 3d 856, 863 (3d Dist. 2011) (cosmetologist not strictly liable for applying permanent eye makeup that was defective); Hernandezcueva v. E.F. Brady Co., Inc., 243 Cal.App.4th 249, 196 Cal. Rptr. 3d 594, 601 (2d Dist. 2015), as modified (Jan. 15, 2016) (analyzing issue of whether a subcontractor that installed defective drywall had a significant “role in the stream of commerce” relating to the product, such that he could be liable). For example, in Ferrari, the issue was whether a white-water rafting company that supplied the plaintiff with a defective raft could be strictly liable. These cases would be on point if a plaintiff got in an Uber ride, was injured because the driver's car was defective, and then tried to hold Uber or the Uber driver strictly liable for the defect in the car. In that case, it would make sense to say that the provision of the product was merely incidental to the provision of the service—the ride. But no one would doubt that the car manufacturer could be held liable for the defect, even though the car was used to effectuate the ride, a service. See Ferrari, 38 Cal. Rptr. 2d at 71 (“For example, an airline passenger injured because of a defect in the craft would have a strict liability claim against the manufacturer․ The manufacturer's role is that of a provider of a product, the airplane. On the other hand, the airline operating the plane would be primarily involved in providing a service, i.e., transportation. The airline is itself the end user of the product and imposition of strict liability would be inappropriate.”). In short, what these “predominant purpose” cases do not say is that any product used for obtaining services, or in relation to a transaction for services, is beyond the ambit of products liability law.

Instead, Uber is much more like the defendant in Green v. ADT, LLC, No. 16-CV-02227-LB, 2016 WL 3208483 (N.D. Cal. June 10, 2016), which provided a home security system that was defective. It was not disputed that home security was in some sense a service, but the plaintiff alleged that ADT had also designed and manufactured the devices that were necessary for the system, and that these were defective. On those allegations—and in rejecting an argument based on the same California case law discussed above—the court held that ADT could be liable on a product defect theory. See id. at *3–*4. Uber provides a service, but it also provides a product—the Uber app—over whose design Uber exercises total control, and which is placed into the stream of commerce by Uber. In this way, Uber is unlike the rafting company in Ferrari or the cosmetologist in Hennigan, who had no such role in relation to the defective products at issue. Uber can be liable for defects in the app.¹⁷

3. Whether Plaintiffs Have Alleged Product Defects

So far, the Court has established that the Uber app is a product, and Uber could be liable for defects in the app. But it does not follow from these conclusions that strict liability applies to everything that might go wrong in relation to the use of the app. A more difficult question arises: Are the alleged “defects” really defects in the Uber app, or are they just problems with Uber's services (or with some other aspect of its business model)?

To put the problem another way, Plaintiffs’ defect theory comes close to suggesting that just because a user's interactions with Uber and Uber drivers are facilitated by the Uber app, every injury arising out of those interactions can be traced to a design defect in the Uber app. For example, Plaintiffs list among their alleged design defects Uber's “[f]ail[ure] to offer timely support during unsafe rides” and its “[f]ail[ure] to add appropriate background checks before and after allowing drivers onto the app.” MC ¶ 473. Neither of these has much, if anything, to do with the app as such. Instead, they have to do with Uber's own customer support and hiring practices. It is plausible that these “defects” could have caused injuries sustained by Plaintiffs during Uber rides, and that those rides were procured via the app. But to say that a failure to conduct an adequate background check or its failure to offer timely support are defects in the Uber app simply stretches the concept too far. These are issues that go to the question of whether Uber breached the applicable standard of care as a provider of services, not whether it placed a defective product into the stream of commerce. Most of the alleged defects fall on the wrong side of this line. For example, “[f]ailing to enact and appropriately enforce a zero tolerance policy toward drivers with a history of inappropriate behavior,” MC ¶ 473, is an issue with Uber's conduct as a provider of services, not an issue with the design of its app. “Failing to appropriately monitor rides to detect known patters of sexual assault,” id., including by using Uber's GPS technology, is again a question of Uber's level of care with respect to its services, not with the design or functionality of the app. This distinction, though sometimes fine, is important to maintain. Without it, there would be nothing to prevent Plaintiffs’ negligence claims from being entirely swallowed up into their strict liability claims.

Some of the alleged defects do concern the design of the app itself. For example, Plaintiffs allege that Uber could have designed the app to provide for video or audio monitoring of rides; it could have designed a GPS-based safety alert system to flag route deviations or “excessive time spent with a passenger at the beginning or end of a route;” or it could have provided for biometric scanning for drivers, presumably to ensure that the driver of a given ride was actually the same person who had applied to be an Uber driver. MC ¶ 487. “Under the risk-benefit test, products ‘may be defective if the design embodies an excessive preventable danger.’ ” Hughes v. Apple, Inc., No. 22-CV-07668-VC, ––– F.Supp.3d ––––, 2024 WL 1141751 (N.D. Cal. Mar. 15, 2024) (quoting McCabe v. Am. Honda Motor Co., 100 Cal.App.4th 1111, 123 Cal. Rptr. 2d 303, 310 (2d Dist. 2002)); see also Am. Tobacco Co., Inc. v. Grinnell, 951 S.W.2d 420, 432 (Tex. 1997) (“In determining whether a product is defectively designed, the jury must conclude that the product is unreasonably dangerous as designed, taking into consideration the utility of the product and the risk involved in its use.”).¹⁸ If some of the product liability claims ultimately go forward, Uber will have every chance to prove that the alternative designs would not address the problems with sexual assault, or that the risk of injury by sexual assault is not “excessive” given the costs of the design, and so forth. At this stage, there is nothing unusual about the coexistence of negligence allegations and products liability allegations in relation to the same alleged injuries. See, e.g., Hughes, ––– F.Supp.3d ––––, 2024 WL 1141751 (denying motion to dismiss negligence-based and design-defect claims regarding Apple's failure to design additional safety features into the AirTag to prevent stalking incidents).

4. Causation

Ultimately, Plaintiffs’ products liability claims founder on the absence of individual allegations that make the causation allegations plausible. As Uber points out, whether a given design defect is a proximate or even but-for cause of any plaintiff's injury will depend on the circumstances of the alleged assault or harassment. For example, if a plaintiff was assaulted by someone of the same gender—or would not have selected a driver of the same gender even if able to do so—then Uber's failure to design that feature into the app did not cause their injury. If a plaintiff was assaulted on a ride without any route deviations or unexpected driving, then designing the app to flag such conduct could not be said to have caused a given assault. And if the alleged assailant was a registered driver, such that any biometric screening would not have prevented him from using the app, then failing to add hat feature did not cause the assault. Here, neither the MC nor the short-form complaints contain any descriptions of the alleged incidents, nor do they otherwise explain how the absence of a given feature caused any particular assault. At least given the parties’ and the Court's understanding about how motions to dismiss were adjudicated, there need to be some allegations making causation plausible not just in a general, abstract sense, but with respect to the claims of the actual plaintiffs in this MDL. The products liability claims are therefore dismissed with leave to amend.

H. Injunctive Relief

Plaintiffs ask the Court for injunctive relief under the UCL. Because injunctive relief targets future harm, the Court must consider whether Plaintiffs have standing. To establish standing, a plaintiff must show (1) that they have suffered an injury in fact, (2) that their injury is fairly traceable to a defendant's conduct, and (3) that their injury would likely be redressed by a favorable decision. Lujan v. Defs. of Wildlife, 504 U.S. 555, 560–61, 112 S.Ct. 2130, 119 L.Ed.2d 351 (1992). Past wrongs, by themselves, are insufficient to establish an injury-in-fact. See Davidson v. Kimberly-Clark Corp., 889 F.3d 956, 967 (9th Cir. 2018). Rather, a plaintiff must show “that he is realistically threatened by a repetition of [the injury].” City of Los Angeles v. Lyons, 461 U.S. 95, 109, 103 S.Ct. 1660, 75 L.Ed.2d 675 (1983)). A plaintiff must plead a “threat of injury [that is] ‘actual and imminent, not conjectural or hypothetical.’ ” Davidson, 889 F.3d at 967 (quoting Summers v. Earth Island Inst., 555 U.S. 488, 493, 129 S.Ct. 1142, 173 L.Ed.2d 1 (2009)).

Plaintiffs have put forward two theories for their injury-in-fact. One focuses on Uber's conduct “deter[ring] Plaintiffs from using the product and thus restrict[ing] their access to transportation.” MC ¶ 358. The other on the “substantial risk” of future sexual assault Plaintiffs face by using Uber. MC ¶ 354. The Court finds that neither theory suffices to establish standing for injunctive relief.

Plaintiffs’ deterrence injury theory is rooted in Davidson. The plaintiff in that case asserted she had standing to request injunctive relief to stop Kimberly-Clark, a manufacturer of pre-moistened wipes, from falsely advertising their wipes as flushable. Davidson, 889 F.3d at 966. The court agreed, finding a “threat of imminent or actual harm by not being able to rely on Kimberly-Clark's labels in the future.” Id. at 967. Plaintiffs here argue that their injury is the same: that they cannot rely on Uber's advertisements of their product as safe, so are harmed because they want to purchase the product as advertised but cannot do so.

But Plaintiffs face the same pleading issues here as they do with the fraud-related claims that the Court addressed above. Plaintiffs have made no showing that any particular plaintiff saw any particular Uber advertisement, let alone that they relied on the representation to purchase the product or that they will forego riding in an Uber in the future because they cannot trust the advertising. In other words, Plaintiffs cannot point to an injury-in-fact relating to Uber's misrepresentations because they have not plead specific facts showing those misrepresentations occurred. A bare allegation that Plaintiffs are unable to purchase Uber's product as advertised does not establish an injury. Without allegations supporting the existence of an injury in the first place, the Court cannot conclude that the threat of that injury recurring is “actual and imminent.” See Davidson, 889 F.3d at 967. To the extent that Plaintiffs seek for the Court to create new law and expand the reasoning of Davidson to apply to a plaintiff who was not falsely advertised to, the Court declines to do so.

Plaintiffs’ “substantial risk” theory of injury is a more traditional future injury—that Plaintiffs may be sexually assaulted again if they ride Uber. Courts have used several methods to assess a risk of recurrence.¹⁹ One is the method used by the Ninth Circuit in Payne v. Office of the Commissioner of Baseball. 705 F. App'x 654 (9th Cir. 2017). There, the court concluded it was appropriate to dismiss a claim for injunctive relief where the likelihood of future injury was 0.0027%, or roughly 3 in 100,000. Id. at 655. Here, Plaintiffs’ allegations do not allow the Court to calculate the exact likelihood that an Uber rider will be sexually assaulted in the future, but numeric precision is unnecessary to conclude the likelihood of injury here is similarly low. For example, the Master Complaint mentions that Uber charged riders for “hundreds of millions of rides,” and quotes Uber's Apple app store listing as inviting “the public to ‘join the millions of riders who trust Uber for their everyday travel needs.’ ” MC ¶¶ 204, 210. Among those millions of rides, Plaintiffs identify roughly 6,000 reports of sexual assault in the years 2017 and 2018. MC ¶ 273. Using those rough numbers as a guide to the per-ride probability that a driver will sexually assault a rider, the Court cannot find that the likelihood is high enough to clear the “actual and imminent” standard of Lyons. Even assuming as true that some plaintiffs will use Uber again (whether they are required to use Uber because they are clients to third parties who use Uber or otherwise), Plaintiffs have made insufficient allegations that any individual plaintiff has more than a “tenuous likelihood of future injury.” Doe v. Match.com, 789 F. Supp. 2d 1197, 1200 (C.D. Cal. 2011). Thus, the Court finds that the “substantial risk” Plaintiffs allege is unable to satisfy Lyons.

In a parenthetical, Plaintiffs obliquely suggest another rationale for calculating the risk of future injury, relying on Armstrong v. Davis. There, the Ninth Circuit upheld a district court's finding that a class of disabled prisoners and parolees had standing to seek an injunction against the California Board of Prison Terms for failing to make proper accommodations under the Americans with Disabilities Act. Armstrong v. Davis, 275 F.3d 849, 854 (9th Cir. 2001). The court described “at least two ways in which to demonstrate that [an] injury is likely to recur,” and thus establish standing. Id. at 861. First, “that the defendant had, at the time of the injury, a written policy, and that the injury ‘stems from’ that policy.” Id. (citing Hawkins v. Comparet–Cassani, 251 F.3d 1230 (9th Cir. 2001)). Or second, “that the harm is part of a ‘pattern of officially sanctioned ․ behavior, violative of the plaintiffs’ [federal] rights.’ ” Id. (citing LaDuke v. Nelson, 762 F.2d 1318 (9th Cir. 1985)). The Armstrong court found that the defendants’ written policy did “not comply with the requirements of the ADA,” which established standing under the “written policy” theory, and, separately, that the Board “routinely deprives disabled prisoners and parolees of their rights under the ADA,” which established standing under the “pattern of officially sanctioned behavior” theory. Id. at 863–4.

Plaintiffs do not argue in their Opposition that Uber had a written policy that caused their injuries. But they do make general allegations about Uber's “official policy” that led to an increased risk of sexual assault on the platform, and one could argue, even if Plaintiffs do not, that Uber's decision to use name-based background checks was effectively a written policy. See MC ¶ 142–145. That argument, however, falls short of the standard set in Armstrong. There, the court reviewed the defendants’ written policy documents and identified specific, textual shortcomings in complying with the ADA. Importantly, the written policy, itself, was violative of the ADA—if the policy was followed, the injury would necessarily occur. Here, even assuming that the decision to pursue name-based background checks was a written policy under Armstrong, the policy itself is not violative of any law nor does it necessarily lead to the injuries Plaintiffs allege. Rather, the policy could increase the chance that a rider might be sexually assaulted in an Uber by allowing poorly screened drivers onto the platform. Plaintiffs would still have to show how their injury stemmed from that inadequate background check. In other words, it would bring the question back to the very one asked in Lyons: how likely is it that the plaintiff will be harmed in the future? A finding that Plaintiffs’ injuries “stem from” a written policy like this one would create an end-run around Lyons that the Court cannot endorse. Accordingly, Plaintiffs cannot establish standing on a written policy injury.

Plaintiffs also argue that “sexual assault is a well-accepted part of Uber's business model,” which could establish standing as a “pattern of officially sanctioned behavior.” Plaintiffs cite to Doe v. Hagee in support for these arguments. The plaintiffs in Hagee sought an injunction against the U.S. Marine Corps after they were sexually assaulted by Marine Corps recruiters. 473 F. Supp. 2d 989, 992 (N.D. Cal. 2007). The Hagee court found that the plaintiffs were entitled to seek an injunction because the “defendants ha[d] not shown that it would be impossible” for the plaintiffs to present evidence that the Marines officially “condoned the practice of sexually assaulting recruit[s].” Id.

On its face, the Hagee decision supports Plaintiffs’ arguments. But the Court notes that it applies an outdated legal standard for adjudicating motions to dismiss. Under Twombly and Iqbal, the burden is not on defendants to show that plaintiffs have alleged impossibilities, but on plaintiffs to make facially plausible allegations. Plaintiffs make no arguments in their Opposition that Uber sanctioned, condoned, or otherwise approved of any sexual assaults of its riders by its drivers, merely stating that sexual assault is a “well-accepted part of Uber's business model” and “that Uber avoids other safety mechanisms, such as cameras, to maximize revenue.” Opp'n at 113. The Court cannot find that those general statements plausibly establish that Uber has engaged in a pattern of officially sanctioning sexual assault. With that being the case, the Court finds that Plaintiffs have not adequately plead their request for injunctive relief.

As Plaintiffs only request injunctive relief for their UCL claims, the Court's decision on Plaintiffs’ standing for injunctive relief renders those claims moot.²⁰ Accordingly, Plaintiffs UCL claims are dismissed.

I. Punitive Damages

Defendants ask the Court to dismiss Plaintiffs’ request for punitive damages. The substantive law of California and Texas regarding punitive damages differs slightly, as discussed below, but the procedural law on which the pleadings are judged is the same. That is, the Court must examine the sufficiency of the pleadings subject to federal standards. See Clark v. State Farm Mut. Auto. Ins. Co., 231 F.R.D. 405, 406 (C.D. Cal. 2005). Under Fed. Rule. Civ. P. 9(b), “malice, intent, knowledge, and other conditions of the mind may be alleged generally.” “[A] plaintiff may include a short and plain prayer for punitive damages that relies entirely on unsupported and conclusory averments of malice or fraudulent intent.” Rees v. PNC Bank, N.A., 308 F.R.D. 266, 273 (N.D. Cal. 2015).

1. California

Under California law, punitive damages are available where a plaintiff proves “by clear and convincing evidence that the defendant has been guilty of oppression, fraud, or malice.” Cal. Civil Code. § 3294(a). Here, Plaintiffs allege that “Uber acted intentionally, recklessly, wantonly, maliciously, and fraudulently.” MC ¶ 341. That allegation, combined with Plaintiffs’ surviving vicarious liability and common carrier's non-delegable duty claims, suffices to state a request for punitive damages. See Sidhu v. Bayer Healthcare Pharms. Inc., No. 22-CV-01603-BLF, 2023 WL 6541865, at *13 (N.D. Cal. Oct. 5, 2023).

2. Texas

Under Texas law, punitive damages are available where a plaintiff proves “the defendant acted with fraud, malice, or gross neglect.” Tex. Civ. Prac. & Rem. Code. § 41.003. As all of Plaintiffs’ Texas law claims against Uber except for the core negligence claims have been dismissed, the only question is whether Plaintiff's allegations rise to the level of gross negligence. As described in Part III(B)(1), above, Texas law has two requirements for gross negligence: “(1) viewed objectively from the standpoint of the actor, the act or omission must involve an extreme degree of risk, considering the probability and magnitude of the potential harm to others, and (2) the actor must have actual, subjective awareness of the risk involved, but nevertheless proceed with conscious indifference to the rights, safety, or welfare of others.” Langston v. Ethicon Inc., No. 3:20-CV-3712-S, 2021 WL 6198218, at *6 (N.D. Tex. Dec. 31, 2021) (quoting Great Plains Tr. Co. v. Morgan Stanley Dean Witter & Co., 313 F.3d 305, 314 (5th Cir. 2002)). Under the objective prong, Texas law “requires [plaintiffs] to prove not only that the defendants’ omissions created a risk of serious harm, but also that the likelihood of the harm occurring was more than remote.” Kovaly v. Wal-Mart Stores Texas, LLC, 157 F. Supp. 3d 666, 670 (S.D. Tex. 2016) (quoting Henderson v. Norfolk S. Corp., 55 F.3d 1066, 1070–71 (5th Cir. 1995)).

Plaintiffs’ pleadings clearly satisfy the subjective prong, as they are replete with allegations that Uber knew of the risk of sexual assault and intentionally did not take action to prevent it. See, e.g., MC ¶ 342–344. The objective prong is somewhat more challenging. It is evident that a sexual assault is a “serious harm.” The question, then, is whether the likelihood of sexual assault occurring was “more than remote.” As discussed above regarding injunctive relief, it is difficult to quantify the likelihood of sexual assault on Uber's platform. But the likelihood analysis for standing, which is focused on a plaintiff's individual risk of injury, is different than that for gross negligence, which is focused on a defendant's actions or omissions. In other words, the question is whether Uber's omissions created a sufficient likelihood of sexual assault on the platform as a whole, not of an individual plaintiff being sexually assaulted again. For at least those drivers who had multiple allegations of sexual assault and whom Uber continued to let drive on its platform, Plaintiffs have alleged that likelihood is not remote. That is, given the magnitude of the harm of a sexual assault, Uber's decision to not remove those drivers created an extreme degree of risk for its users. Therefore, at the pleading stage, Plaintiffs’ allegations suffice to keep alive their request for punitive damages under Texas law.

J. Choice of Law

Finally, in its Texas Motion, Uber asks the Court to conduct a choice-of-law analysis to determine what law applies to which plaintiffs. In support, it cites several MDL cases in which the choice-of-law was determined at the motion to dismiss stage and argues resolving the question “is consistent with this Court's MDL role.” Notably, Uber cites to no authority requiring that the Court decide the question nor is the Court aware of any such authority.

The Court declines to reach the question at this stage. Since Uber did not move to dismiss all of Plaintiffs’ claims under any state's law, determining choice-of-law will not be dispositive of any individual plaintiffs. And although Uber asks the Court to hold Plaintiffs’ decision to not completely brief the choice-of-law issue against them, the Court is not convinced that deciding the issue at this juncture on incomplete briefing is the appropriate course. Further, as Plaintiffs point out, some of the analysis will likely turn on the outcome of Uber's appeal of the Court's order on its Terms of Use Motion, which still pends. The Court will reserve the question for a later date and will likely request more fulsome briefing of the factual specifics, including on Plaintiffs’ suggestion of disagreement with Uber's assertion of which plaintiffs assented to which Terms of Use.

IV. CONCLUSION

For the foregoing reasons, the Court orders as follows:

Defendants’ Motion to Dismiss Claims C, D, F, H, and I under California law is GRANTED. Defendant's Motion to Dismiss Claim G under California law is DENIED.

Defendants’ Motion to Dismiss Claims C, D, E, F, G, and H under Texas law is GRANTED.

Dismissal is with leave to amend. A CMC is set for August 29, 2024, at 10:00 a.m., at which time the Court and the parties will discuss the timeline and procedures for the filing, and challenging, of any amended pleadings. The parties shall file a joint status statement by August 27.

IT IS SO ORDERED.

FOOTNOTES

1.   To be clear, all of the claims involve several common issues of fact, and Uber's arguments tend to acknowledge this in most cases. (And notably, Uber also has not moved to dismiss the plaintiffs’ core negligence claims in any case.) The Court can and will adjudicate the arguments that do not require individualized allegations. It is just that certain elements of some claims—reliance, for example, or causation for certain claims—tend also to require individual allegations if they are to be plausible alleged.

2.   It is true that some of the necessary facts are exclusively within Uber's knowledge—for example, Uber alone knows what its background checks turned up and whether it reinstated or declined to discipline drivers accused of assault. But plenty of case-specific facts are within Plaintiffs’ knowledge: how and when assaults took place, whether an assault was reported and to whom, what safety-related marketing plaintiffs saw, etc.

3.   See Part III(J) below.

4.   Uber does not argue that the Texas TNC Statute has any retroactive effect. See Tex. Mot. at 6.

5.   The Court parts ways from the reasoning in Doe v. Uber Techs., Inc., No. 19-CV-03310-JSC, 2019 WL 6251189, at *4–*5 (N.D. Cal. Nov. 22, 2019), on the question of foreseeability. First, the parties do not seem to have brought the analysis from cases like John R. and Mary M. (which focus almost exclusively on policy and foreseeability) to the court's attention. Second, the court there discounted any discussion of the nature of Uber's business model as simply another way of talking about the plaintiff's vulnerability. But there is a difference between discussing the “conditions” of work and the nature of the “enterprise”—both well-established aspects of the foreseeability inquiry—and emphasizing the plaintiff's vulnerability.

6.   As an aside, the Court finds the “motivating emotions” test to be extremely puzzling. If taken at face value, the test would make it impossible to conceive of any situation in which a sexual assault would fall within the scope of employment (a result that would be at odds with the Supreme Court of California's explicit decision not to create a per se rule to that effect). More troublingly, the “motivating emotions” inquiry seems ultimately to depend on abstract, unarticulated, and probably unfounded psychological assumptions about what sorts of “conditions” motivate sexual assailants to feel the “emotions” that drive them to commit crimes. Drawing a distinction between assaults that arise out of the conditions of work and those that arise out of mere “propinquity and lust” doesn't help, either. See Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 364. It seems axiomatic that all or nearly all sexual assaults arise out of “propinquity and lust.” At least, it is not clear why—to take Lisa M.’s example—a therapist who assaults a patient with whom he has a long-term professional relationship should be regarded as less motivated by “propinquity and lust” than, say, an Uber driver who assaults a passenger. Is the assumption supposed to be that having a longer-term relationship with someone makes it more likely that the “motivating emotions” to sexually assault them will arise? The idea seems strange, particularly now, more than 30 years on, at a time when the social understanding of sexual assault and sexual harassment have been significantly transformed.

7.   This result would also follow from the rule applied in Xue Lu. “Although a circuit court's prediction of state law is not binding in the same way as is its definitive interpretation of federal law, as a practical matter a circuit court's interpretations of state law must be accorded great deference by district courts within the circuit.” Johnson v. Symantec Corp., 58 F. Supp. 2d 1107, 1111 (N.D. Cal. 1999) (Fogel, J.).

8.   While the cases contain some confusion about the distinction between apparent agency and apparent authority, the approach of California and Texas appears to be to treat the distinction as unimportant. See Baptist Mem'l Hosp., 969 S.W.2d at 947 n.2 (“Many courts use the terms ostensible agency, apparent agency, apparent authority, and agency by estoppel interchangeably. As a practical matter, there is no distinction among them.”); see also Kaplan, 59 Cal. App. 4th at 747–48, 69 Cal.Rptr.2d 640 (discussing apparent authority and ostensible agency apparently interchangeably). Plaintiffs endorse this view in their Opposition. See Opp'n at 57. It is also the modern approach of the Restatement (Third) of Agency, § 2.03. See also id. § 2.03, cmt. a.

9.   There are some Texas cases that are at least facially at odds with this analysis. For example, in Wyndham Hotel Co. v. Self, 893 S.W.2d 630 (Tex. App.--Corpus Christi 1994), writ denied (Oct. 5, 1995), the court said that “a cause of action based on agency by estoppel does not necessarily require a separate finding that the ostensible agent was acting within the scope of apparent authority.” Id. at 634. “Under this formulation of the theory, actions beyond the scope of any hypothetical agency are addressed in terms of whether the party's belief and consequent reliance were reasonable and justified.” Id. But this case seems best understood as essentially in agreement with the Court's analysis here. In other words, there may not be a separate “scope” inquiry, but scope still matters in the sense that one cannot reasonably believe that an agent has unlimited authority just because they seem like an employee.

10.   In Lisa M. itself, the Supreme Court of California noted that the ultrasound technician was actually an independent contractor, but that the defendant had not sought “summary judgment on the ground Tripoli was not its employee, did not argue that issue in the Court of Appeal, and does not rely on it in this court,” so the court simply “treat[ed] Tripoli as Hospital's employee, without considering or deciding whether Tripoli was Hospital's nonemployee agent or ostensible agent.” Lisa M., 48 Cal.Rptr.2d 510, 907 P.2d at 360 n.2. This clearly suggests that the result would have been the same if the technician had been an ostensible agent rather than an employee. In other words, it suggests that the scope of vicarious liability would be no different (or at least no greater) if the assailant had been an ostensible agent rather than an employee.

11.   Plaintiffs’ efforts to get around the Texas TNC statute on this point are unpersuasive. The language is clear and unequivocal, and Plaintiffs ask the Court to read limitations into the statute—such as a distinction between the “regulatory” and “common law” definitions of common carrier—that are just not there. Neither do Durham Transp., Inc. v. Valero, 897 S.W.2d 404 (Tex. App. 1995), or McClure v. Greater San Antonio Transportation Co., 2009 WL 10670097 (W.D. Tex. Mar. 24, 2009), help Plaintiffs. In Durham, the statute at issue “explicitly reserve[d] the issue of the duties and liabilities of carriers for the common law except where otherwise provided in the regulations of carriers.” 897 S.W.2d at 409. In McClure, the defendants’ argument was essentially that a local regulatory scheme had displaced the common law rules regarding common carriers; there was no suggestion that the local ordinance expressly addressed whether the defendant was or was not a common carrier. See 2009 WL 10670097 at *4. That is a far cry from the statutory language here.

12.   If it turned out later that Uber's entrustment of its decals and signs to certain drivers might be relevant to other negligence theories under the laws of either state, Plaintiffs would not be precluded from making that argument.

13.   Cf. S. Cent. Bell Tel. Co. v. Barthelemy, 643 So. 2d 1240, 1246 (La. 1994) (“[S]oftware ․ is physically manifested in machine readable form by arranging electrons, by use of an electric current, to create either a magnetized or unmagnetized space ․ The computer reads the pattern of magnetized and unmagnetized spaces ․ This machine readable language or code is the physical manifestation of the information in binary form.”); cf. also Beyer v. Symantec Corp., 333 F. Supp. 3d 966, 979 (N.D. Cal. 2018) (Chen, J.) (“computer software ․ may be characterized as tangible property” because it “takes up space” and “constitutes a corporeal body”).

14.   The fact that the smartphone itself is a product does not say anything about whether a smartphone application is a product. Both a lamp and a lightbulb are products, and both could be separately defective, even though the former is required for the latter to work.

15.   Electricity, another intangible thing recognized as a product in both California and Texas, actually provides another worthy (if imperfect) analogy. Even though electricity, like the Uber app, is not quite tangible and requires a massive background infrastructure to be useful, courts recognize that it is a product once it enters a consumer's home. See Pierce v. P. Gas & Electric Co., 166 Cal.App.3d 68, 212 Cal. Rptr. 283, 290–91 (3d Dist. 1985); Houston Lighting & Power Co. v. Reynolds, 765 S.W.2d 784, 785 (Tex. 1988). An analogy can be drawn to Uber's software, which may not be a product just in itself, but is sensibly recognized as a product when a user downloads the frontend interface onto their own device.

16.   This conclusion is buttressed by looking to policy considerations underlying strict liability. See Restatement (Third) of Torts: Prods. Liab. § 19, Reporters Notes, cmt. a (“When the applicable definition fails to provide an unequivocal answer, decisions regarding whether a ‘product’ is involved are reached in light of the public policies behind the imposition of strict liability in tort.”). The reasoning of the court in Brookes v. Lyft, Inc., is persuasive: “[T]he designer of ․ the Lyft application placed it into the stream of commerce to derive a profit, is in the best position to protect the user and innocent bystander against the risk of harm associated with any design defect.” 2022 WL 19799628, at *4-5 (Fla. Cir. Ct. Sept. 30, 2022); see also McIndoe v. Huntington Ingalls Inc., 817 F.3d 1170, 1173 (9th Cir. 2016) (“The general aim of strict liability is to place responsibility on the party most able to prevent harm caused by dangerous products and thus to incentivize proper design and quality control of such products. Therefore, strict liability should be imposed on the party best able to protect persons from hazardous equipment.”) (cleaned up).

17.   The California and Texas TNC statutes do not say otherwise. At most, they characterize rideshare rides as services, a point not in dispute. See Cal. Pub. Util. Code § 5431(c); Tex. Occ. Code. §§ 2402.113; 2402.115; 2402.153. To the extent that Uber argues that the statutes show that “the predominant purpose of the Uber App is to facilitate a service, not to sell a tangible product,” Reply at 79, its argument depends on the same misreading of the “predominant purpose” cases discussed above.

18.   As Plaintiffs point out, Uber's briefing seems to suggest that the Uber app could only be defective if it had a coding “glitch” that caused it not to function; but this would be something more like a manufacturing defect than a design defect.

19.   Plaintiffs’ argument that assessing the likelihood of future injury “is a fact question not appropriate for resolution now” is clearly misplaced. The Court may only exercise jurisdiction over actual cases or controversies, which, in the case of injunctive relief, requires an analysis of the likelihood of future harm. See, e.g., Payne v. Office of the Commissioner of Baseball, 705 F. App'x 654, 655 (9th Cir. 2017).

20.   It bears noting that although the Court finds Plaintiffs have not plausibly alleged Article III standing for the purposes of injunctive relief, they likely have satisfied the distinct standing requirements of a UCL claim. For standing under the UCL, a plaintiff must show not only an injury in fact but that they suffered “lost money or property as a result of the unfair competition.” Mastel v. Miniclip SA, 549 F. Supp. 3d 1129, 1144 (E.D. Cal. 2021) (quoting Birdsong v. Apple, Inc., 590 F.3d 955, 959 (9th Cir. 2009)). Plaintiffs have alleged just that. MC ¶ 511.

CHARLES R. BREYER, United States District Judge