A Look at the Legal Environment for Driverless Vehicles (2016)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

30 more reliable and more private than current online ride-service programs in which vehicles come with human drivers. Variations on use of on-demand driverless vehicles appear likely to include coopera- tives in which fleets of driverless vehicles are owned in common by groups and made available for per- sonal or shared use by members of a driverless vehi- cle cooperative. Indeed, driverless vehicles could well become the main type of vehicle used in online on-demand ride services, an increasingly popular form of spon- taneous personal mobility. To the extent that such transportation services replace personal vehicle ownership, such a transformation would reflect fun- damental changes in expectations about personal mobility. Instead of purchasing a machine that requires maintenance and garage space, personal mobility could become a service that requires nei- ther. Some studies of millennials indicate a shift toward such a preference for personal transporta- tion as a service, as opposed to personal mobility through individual vehicle ownership.246 b. Small Low-Speed Driverless Vehicles.—In mid- 2015, Google, Inc., began to take delivery of a fleet of small, low-speed vehicles.247 The company intends to license these two-person vehicles in California, once the California Department of Motor Vehicles has adopted regulations that permit their licensing for on-road operation, enabling general operation of these driverless vehicles on public roadways. Google managers note that the corporation does not plan to go into the business of driverless car manufacturing. Rather, at least initially, the corporation plans to own and use its fleet of 100 or so driverless vehicles to transport employees and visitors on and around the corporation’s campus. For regulatory purposes, these driverless cars are categorized as low-speed vehicles (LSVs) limited to top speeds between 20 and 25 miles per hour. NHTSA regulates these LSVs under a special regulatory category for lighter, lim- ited-speed vehicles, such as golf carts, under Motor Vehicle Safety Standard No. 500, which authorizes small, light vehicles restricted to speeds under 25 miles per hour.248 Low speed driverless vehicles will be limited to use on protected and well-mapped routes. Whether such an application of driverless vehicles can be extrapolated to broader consumer uses remains to be seen. If consumer models of driver- less vehicles are limited to the LSV regulatory category of small, light, low-speed vehicles, the potential consumer market may also be limited to retirement and other planned communities that emphasize alternatives to conventional automo- biles.249 On the other hand, in some large cities, such as New York, the maximum speed limit is already 25 miles per hour.250 In such congested urban cores, small, low-speed, few-passenger driv- erless vehicles would be especially attractive mobility options for local trips. c. Controlled Roadway Environments for Driver- less Vehicles.—In order to minimize legal risks that may result from personal injuries, property damage, or other adverse interactions with unpredictable human drivers and pedestrians, driverless vehicles may work best in controlled environments, in which vehicle movements and roadway events are more predictable. Segregated roadways with few or no pedestrians or other road users, where all vehicles are driverless and where road closures, construc- tion, or repairs are infrequent and mapped in advance, may be the most efficient near-term way to deploy driverless vehicles. In a longer timeframe, probably before the begin- ning of the next century, all vehicles may be required to be driverless. Human-driven vehicles may be pro- hibited from using most public roadways, just as horses are often prohibited on urban streets. IV. CIVIL LIABILITY FOR PERSONAL INJURY For as long as they have been imagined, driverless vehicles have been perceived as a panacea to the safety hazards associated with their human-operated 246 AmeriCAn PUbliC TrAnsPorTATion AssoCiATion, mil- lenniAls & mobiliTy: UndersTAnding The millenniAl mind- seT (2013), http://www.apta.com/resources/reportsand publications/Documents/APTA-Millennials-and-Mobility. pdf. 247 Mark Harris, Google reveals plans to increase pro- duction of self-driving cars, The gUArdiAn (Sept. 12, 2015), http://www.theguardian.com/technology/2015/sep/12/ google-self-driving-cars. 248 Low Speed Vehicles, 49 C.F.R. § 571.500 (2011). 249 See William L. Garrison & David M. Levinson, The TrAnsPorTATion exPerienCe: PoliCy, PlAnning, And dePloy- menT 418 (2014) (describing examples including the Del Web Sun City retirement communities and Celebration in Florida); Daniel Sperling & Deborah Gordon, TWo billion CArs: driving ToWArd sUsTAinAbiliTy 41 (2010). 250 Transcript: Mayor de Blasio Signs New Law Low- ering New York City’s Default Speed Limit to 25 MPH (Oct. 27, 2014), http://www1.nyc.gov/office-of-the-mayor/ news/494-14/transcript-mayor-de-blasio-signs-new-law- lowering-new-york-city-s-default-speed-limit-25-mph (announcing that from November 7, 2014, onward, the speed limit on all streets in New York City will be 25 miles per hour unless otherwise posted); Jim Gorzelany, The World’s Most Traffic-Congested Cities, FORBES (Apr. 25, 2013), http://www.forbes.com/sites/jimgorzelany/2013/04/ 25/the-worlds-most-traffic-congested-cities/ (noting that Manhattan’s actual average road speed hovers around 13 miles per hour).

31 counterparts.251 Yet even if driverless vehicles are safer than other methods of transportation, they will still get into accidents.252 When this happens, ques- tions will arise regarding who should have to bear the costs of these accidents. There exists a well-established body of law that prescribes the legal liabilities of manufacturers and operators of conventional automobiles when these devices injure people or property. This section reviews these rules and considers how they may apply to driverless vehicles.253 These civil liability projections involve substan- tial speculation. It is far from certain that the laws of tomorrow will be the same as those of today. Also, driverless technologies (or the market for these devices) may or may not evolve in the manner pres- ently forecast. These and other contingencies make it difficult to pinpoint how liability rules and driverless vehicles will intersect. Nevertheless, as discussed earlier in this report certain patterns tend to recur when new technologies lead to per- ceived risks of personal injuries. These trends, as mapped against the current law and the antici- pated trajectory of driverless vehicles’ development and diffusion, afford a basis for probabilistic, as opposed to definite, forecasts. A. Basic Principles According to NHTSA, an estimated 5,615,000 police-reported motor vehicle traffic crashes occurred in 2012.254 Although only a small percentage of these crashes generated lawsuits, automobile accidents produce more personal injury lawsuits than do any other type of accident.255 When automobile accidents lead to personal injury and property-damage lawsuits, these cases typically involve application of one or more theories of recovery in tort. Lawsuits against the drivers of automobiles, or against other persons on or about the road, typically sound in a negligence theory of liability. Manufacturers, distributors, and sellers of a vehicle also may be held liable for their negligence, but more often are alleged to have produced a “defec- tive” product under one or more theories of “strict products liability.” The text below summarizes these various avenues of recovery. 1. Negligence Negligence represents the most fundamental and pervasive theory of liability for accidents. A plaintiff suing for negligence must plead and prove a series of “essential elements” in order to recover dam- ages.256 Specifically, the plaintiff must establish that: • The defendant owed the plaintiff a duty of care, with the standard normally being set at “rea- sonable” or “ordinary” care; • The defendant failed to exercise the required care, an element often referred to as “breach”; • The plaintiff’s harm was caused by the defen- dant’s breach of its duty of care (i.e., by the defendant’s negligence)—in other words, had the defendant acted with reasonable care, the harm would not have occurred (an element often referred to as “cause in fact,” or “but-for causation”);257 251 See, e.g., normAn gel geddes, mAgiC moTorWAys 56–57 (1940) (discussing the safety features anticipated to appear in the automobiles of 1960); Martin Mann, The Car That Drives Itself, PoPUlAr sCienCe, May 1958, at 75. 252 Indeed, in May 2015 Google acknowledged that its self- driving vehicles had been involved in 11 minor accidents. These mishaps were attributed to human errors, not the self- driving capabilities of these devices. Justin Pritchard, Google Acknowledges 11 Accidents with Its Self-Driving Cars, sAn Jose merCUry neWs (May 10, 2015), http://www.mercury news.com/california/ci_28091208/business-special-reports. 253 Several commentators already have considered whether and to what extent these principles will apply to the makers and users of driverless vehicles. E.g., RAND rePorT, supra note 181, at 111–34; F. Patrick Hubbard, “Sophisticated Robots”: Balancing Liability, Regulation, and Innovation, 66 flA. l. rev. 1803 (2014); Gary E. Marchant & Rachel A. Lindor, The Coming Collision Between Autonomous Vehicles and the Liability System, 52 sAnTA ClArA l. rev. 1321 (2012); John Villasenor, Prod- ucts Liability and Driverless Cars: Issues and Guiding Principles for Legislation, brookings (Apr. 24, 2014), http://www.brookings.edu/research/papers/2014/04/ products-liability-driverless-cars-villasenor. 254 nATionAl highWAy TrAffiC sAfeTy AdminisTrATion, TrAffiC sAfeTy fACTs: reseArCh noTe 2012 moTor vehiCle CrAshes: overvieW, DOT HS 811 586, 4 (2013). 255 ThomAs h. Cohen, TorT benCh And JUry TriAls in sTATe CoUrTs, 2005 1, U.S. Department of Justice, Bureau of Justice Statistics (Nov. 6. 2009). See also deborAh r. hensler, m. sUsAn mArqUis, AllAn f. AbrAhAmse, sAndrA h. berry, PATriCiA A. ebener, elizAbeTh g. leWis, e. AllAn lind, roberT J. mACCoUn, WillArd g. mAnning, JeAnneTTe A. rogoWski & mAry e. vAinA, ComPensATion for ACCiden- TAl inJUries in The UniTed sTATes 122–23 (1991) (describing the linkage between automobile accidents and “claiming” behavior by injured parties). 256 dAn b. dobbs, The lAW of TorTs § 125 (2d ed. 2011). 257 This inquiry normally requires a counterfactual in which all other circumstances are held constant, except that the allegedly negligent party acted with the requisite care. If the same harm would have transpired anyway notwith- standing these changed circumstances, the requisite causal connection does not exist. resTATemenT (Third) of TorTs: li- AbiliTy for PhysiCAl And emoTionAl hArm § 26, cmt. (2010). In limited circumstances, a party’s obligation to prove but-for causation (also sometimes referred to as “cause in fact”) will be relaxed. This situation sometimes arises when defendants whose negligence already has been established occupy a su- perior position, relative to the plaintiff, to prove that their breach of a duty did not contribute to the plaintiff’s injuries. In chain collisions involving seriatim accidents among auto- mobiles, for example, some courts shift the burden to mul- tiple negligent defendants to establish that their negligence was not the cause of all or part of the injuries incurred by a plaintiff buffeted by multiple impacts. E.g., Fugere v. Pierce, 5 Wash. App. 592, 597, 490 P.2d 132, 135–36 (Wash. App. 1971).

32 • The defendant’s negligence also must represent a “proximate cause” of the plaintiff’s harm, meaning that a reasonable person in the defendant’s position at the time of his or her claimed negligence would have fore- seen that their negligent behavior could lead to harm of the general sort suffered by the plaintiff;258 and • Damages.259 The duty element of a negligence claim is a “matter of law,” meaning that judges, rather than juries, ordi- narily ascertain whether a defendant owed a plaintiff a duty.260 With conventional automobile accidents, the existence of a duty is only rarely a disputed issue.261 Matters that are contested more often concern whether the defendant breached his or her duty of reasonable care, whether this breach caused the plaintiff’s injuries, the extent of the plaintiff’s dam- ages, and whether these damages are wholly or in part due to the plaintiff’s own negligence. The last of these matters represents a defense commonly referred to as “comparative negligence” or “comparative fault.”262 Unlike questions of duty, these issues are left to juries to decide, unless the parties have stipulated to a trial before a judge, or a judge determines that given the facts involved, all reasonable juries would have to agree how an issue should be decided. In evaluating the viability of a negligence claim resulting from a vehicle accident, parties, attorneys, and judges benefit from a robust body of case law that has accumulated over the past 115 years.263 Opinions issued by judges in earlier cases and records of jury verdicts help lawsuit participants anticipate the reso- lution of disputes. Judicial opinions, in particular, provide guidance in the application of various rules associated with the presentation of vehicle accident claims that sound in negligence. Such topics include whether a statute, regulation, or ordinance will flesh out the generic standard of reasonable care, under a doctrine known as negligence per se;264 and whether circumstantial evidence permits a jury to infer a par- ty’s negligence, under a doctrine known as res ipsa loquitur.265 The predictability of the outcomes in many automobile-accident cases has led to the “routi- nization” of a substantial segment of legal practice in this field. High-volume “settlement mills” resolve many of these matters, especially those that involve only modest damages.266 2. Strict Products Liability Negligence law also applies to the manufacturers and sellers of automobiles.267 Furthermore, in the vast majority of states these parties also may be sued under an alternative theory of tort liability. This form of liability does not rest squarely on notions of “fault,” as does negligence. Instead, this approach holds product manufacturers and other defendants closely involved in a product’s chain of distribution and sale liable without a showing of fault for “defects” in the products they make, distribute, or sell.268 258 Proximate causation conventionally has been described in terms of whether a type of harm was fore- seeable at the time of the negligent act or omission, or as an inquiry into whether an actor’s negligence was a “substantial factor” in bringing about the claimed harm. resTATemenT (Third) of TorTs: liAbiliTy for PhysiCAl And emoTionAl hArm, supra note 257, at § 29 cmts. a, e. More recently, a formulation has emerged whereby proximate cause relates to the “scope” of the enhanced risk created by an actor’s negligent behavior. The resTATemenT (Third) of TorTs: PhysiCAl And emoTionAl hArm, an advocate of this approach, provides that “[a]n actor’s liability is lim- ited to those harms that result from the risks that made the actor’s conduct tortious.” Id. § 29. 259 dobbs, supra note 256, at § 124. 260 resTATemenT (Third) of TorTs: liAbiliTy for PhysiCAl And emoTionAl hArm, supra note 257, § 7 cmt. b. 261 60A CorPUs JUris seCUndUm moTor vehiCles § 582 (2015) (“The operator of a motor vehicle has a duty to exercise reasonable or ordinary care for the safety of others while operating his vehicle.”) 262 dobbs, supra note 256, at § 220. 263 The first published decision in a tort lawsuit involv- ing an automobile was Mason v. West, 31 Misc. 583, 65 N.Y.S. 651 (C.C.N.Y. 1900), rev’d, 70 N.Y.S. 478 (N.Y.A.D. 1901). 264 resTATemenT (Third) of TorTs: PhysiCAl And emo- TionAl hArm, supra note 257, § 14 (“An actor is negligent if, without excuse, the actor violates a statute that is designed to protect against the type of accident the actor’s conduct causes, and if the accident victim is within the class of persons the statute is designed to protect.”). 265 Id. § 17 (“The factfinder may infer that the defen- dant has been negligent when the accident causing the plaintiff ’s harm is a type of accident that ordinarily hap- pens as a result of the negligence of a class of actors of which the defendant is the relevant member.”) 266 Nora Freeman Engstrom, Run-of-the-Mill Justice, 22 geo. J. leg. eThiCs 1485, 1500 (2009) (discussing the emergence of “settlement mills” for soft-tissue automobile accident claims). 267 dAvid g. oWen, ProdUCTs liAbiliTy lAW §§ 2.1–2.6 (2008). 268 Warranty law also provides a means for purchasers of goods to recover against the seller, regardless of negli- gence, when the purchased products proved defective. To- day, the few states that have not adopted strict products liability in tort tend to recognize similar rights of redress cast as warranty protections. Graham, supra note 124, at 620. In jurisdictions that recognize strict products liability claims, a cause of action alleging a breach of a warranty may provide an additional basis for recovery. The most important warranty associated with the sale of goods by merchants is the implied warranty of merchantability. Under Section 2-314 of the Uniform Commercial Code, “a warranty that the goods shall be merchantable is implied in a contract for their sale if the seller is a merchant with respect to goods of that kind.” U.C.C. § 2-314. To be “merchantable,” fungible goods must be “of fair average quality” and “fit for the ordinary purposes for which such goods are used.” Id.

33 There exist three such forms of “strict products liability,” which are distinguished from one another by the nature of the defect that the product is alleged to contain: • Manufacturing Defects. A product contains a “manufacturing defect” when it “departs from its intended design even though all possible care was exercised in the preparation and marketing of the product.”269 This sort of defect applies to products that are “physically flawed, damaged, or incorrect- ly assembled,”270 such as a soda bottle cast with glass less thick than the manufacturer specified, making the bottle prone to break. The prevalence of products with this sort of defect in the 1940s and 1950s helped inspire the trend toward strict prod- ucts liability.271 Due to modern quality-control measures, however, claims that allege manufac- turing defects are now few and far between.272 • Design Defects. By contrast, design defect claims, once rare, are now relatively common.273 There exist two approaches toward recognizing a defect in a product’s intended design. One, known as the “consumer expectations” or “consumer con- templation” standard, regards a product’s design as defective when it fails to perform as safely as an ordinary consumer would expect under the circum- stances presented.274 An alternative approach is sometimes referred to as the “risk-utility” definition of a design defect. Per one leading authority that en- dorses this view of a design defect, a product is defective in design when the foreseeable risks of harm posed by the product could have been reduced or avoided by the adoption of a reasonable alternative design by the seller or other distributor, or a predecessor in the commercial chain of distribution, and the omission of the alternative design renders the product not reasonably safe.275 Today, few jurisdictions apply only the “consumer expectations” definition of a design defect.276 Most states have adopted some form of the risk-utility test as the principal means for ascertaining a design defect. Nevertheless, in evaluating whether a par- ticular product’s design is “not reasonably safe,” these jurisdictions commonly encourage the consid- eration of a wide range of factors that may include the safety expectations of consumers.277 Other juris- dictions utilize a hybrid approach that more directly implicates both the consumer expectations standard and a risk-utility formulation.278 • Warning Defects. Third and finally, a product is regarded as defective when a failure to provide sufficient instructions or warnings renders the product “not reasonably safe” in light of the “fore- seeable risks of harm posed by the product.”279 This standard, with its emphasis upon the need for rea- sonable notice of foreseeable risks, deviates only modestly from a negligence rule. Many judicial opinions have elaborated how these principles apply to automobiles.280 It is today well accepted that automobiles must be designed in a manner that makes them reasonably safe even in the event of an accident—in other words, they must be “crashworthy.”281 It is also generally accepted that a plaintiff ’s negligence can represent a full or par- tial defense in a products lawsuit.282 However, there exists no similar consensus about other issues asso- ciated with products liability cases, such as whether a plaintiff will be regarded as having breached his or her duty of exercising reasonable care for their own safety by failing to utilize a seat belt.283 Personal injury lawsuits against manufacturers and sellers of driverless vehicles —whether framed in negligence, strict liability, or other theories— likely will draw to some degree from decisions in prior cases involving products such as conventional vehicles and their components, GPS devices,284 269 23 C.F.R. § 636.109. 270 Id. § 2 cmt. c. 271 Graham, supra note 124, at 600–13. 272 Aaron D. Twerski, Chasing the Illusory Pot of Gold at the End of the Rainbow: Negligence and Strict Liability in Design Defect Litigation, 90 mArq. l. rev. 7, 18 (2006) (“Manufacturing defects are rare events.”) 273 JAne sTAPleTon, ProdUCT liAbiliTy 30 (1994). 274 Mikolajczyk v. Ford Motor Co., 231 Ill. 2d 516, 526– 27, 901 N.E.2d 329, 336 (2008). 275 resTATemenT (Third) of TorTs: ProdUCTs liAbiliTy, supra note 269, § 2. 276 Id., § 2 cmt. d. 277 Id., § 2 cmt. f. See also John W. Wade, On the Nature of Strict Tort Liability for Products, 44 miss. L.J. 825, 837– 38 (1973) (relating factors relevant to the design-defect inquiry). 278 See, e.g., Soule v. Gen. Motors Corp., 8 Cal. 4th 548, 566–68, 882 P. 2d 298, 307-309 (1994); Mikolajczyk, 231 Ill. 2d at 555–57 (Ill. 2008). 279 resTATemenT (Third) of TorTs: ProdUCTs liAbiliTy supra note 269, § 2. 280 This type of litigation is so common as to merit a full chapter in the leading single-volume products liability treatise. oWen, supra note 267, §§ 17.1–17.5. 281 resTATemenT (Third) of TorTs: ProdUCTs liAbiliTy, supra note 269, § 16 & cmt. a. 282 Id., § 17 & cmt. a. 283 dobbs, supra note 256, § 231. 284 E.g., American Winds Flight Acad. v. Garmin Intern., 2010 U.S. Dist. LEXIS 97651 (N.D. Ohio Sept. 17, 2010) (holding that the manufacturer of a GPS device was not liable for the death of a pilot in an airplane equipped with the device, on grounds including the open and obvious nature of the danger associated with exclusive reliance upon the device).

34 autopilot functions on airplanes,285 and aeronautical charts.286 As to certain issues, these cases have pro- duced a robust body of seemingly pertinent case law; as to other topics, there are fewer decisions on point. Furthermore, the risks and benefits of driverless vehicles may be sufficiently distinctive that these decisions may not provide useful references in all circumstances. 3. Limitation or Preemption of Tort Liability Critics of the tort system sometimes express con- cerns that the application of common-law tort prin- ciples through the courts can involve excessive costs, insufficient or poorly allocated benefits, and the rec- ognition of conflicting or suboptimal standards of conduct. An additional fear is that even the mere threat of tort liability potentially can frustrate the development and diffusion of socially beneficial technologies. These concerns have prompted efforts to avoid or limit conventional tort law in particular contexts: • No-Fault Insurance Laws. In the 1960s and 1970s, many states enacted “no-fault” automobile insurance regimes.287 Under these laws, claims that allege only minor property damage or modest personal injuries bypass the tort system altogether, and are instead compensated by first-party insur- ance on a “no-fault” basis.288 It was hoped that “no-fault” regimes would result in more compen- sation for accident victims at a cost savings rela- tive to conventional systems that depend more heavily upon courtroom litigation for recovery.289 After a promising start, the no-fault movement stalled in the mid-1970s due to circumstances such as opposition among plaintiffs’ lawyers, dis- appointing early results that failed to meet the expectations of no-fault advocates, and the closing of a “policy window” conducive to the enactment of no-fault schemes.290 These dynamics have led a few states to repeal the no-fault statutes they had enacted.291 • Liability Caps. The federal government and many states have placed caps on the damages that claimants can recover in certain contexts.292 One type of cap limits the damages that can be received for a particular type of harm in a class of accidents or lawsuits, such as the ceilings that some states have placed on noneconomic damages (such as pain and suffering) in medical malpractice cases.293 • Alternative Forums. Statutes also can channel claims toward forums other than conventional courts. For example, in response to concerns that liability pressures were depleting the supply of childhood vaccines, Congress enacted the National Childhood Vaccine Injury Act of 1986.294 This law directs personal injury claims associated with the administration of a listed vaccine toward a special federal forum, in which cases are heard by special masters, rather than juries.295 These masters can award damages (funded by an excise tax on vac- cines), but the pain and suffering damages that can be awarded for successful claims are capped at $250,000.296 • Preemption by Statute: Congress and state leg- islatures can eliminate tort liability by preempting these claims.297 One rationale for preemption involves the protection of an emerging industry from the threat of future litigation. In this spirit, the Biomaterials Access Assurance Act of 1998 con- ferred broad immunity upon the suppliers of bio- materials for injuries associated with implants that incorporated these biomaterials, but were manufactured by other parties.298 Likewise, by pro- viding that “[n]o provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider,” Section 230 of the 285 Brouse v. United States, 83 F. Supp. 373, 374 (N.D. Ohio 1949). See also Dylan LeValley, Autonomous Vehicle Liabil- ity—Application of Common Carrier Liability, 36 seATTle Univ. l. rev. supra 5, 9–10 (2013) (critiquing the analogy between autonomous vehicles and autopilot functions). 286 See generally David L. Abney, Liability for Defective Aeronautical Charts, 52 J. Air l. & Com. 323 (1986). 287 JAmes m. Anderson, PAUl heATon & sTePhen J. CArroll, The U.s. exPerienCe WiTh no-fAUlT AUTomobile insUrAnCe 35–44 (2010). Engstrom, supra note 103, at 303. 288 Engstrom, supra note 103, at 320–22. 289 See generally roberT e. keeTon & Jeffrey o’Connell, bAsiC ProTeCTion for The TrAffiC viCTim (1965). 290 Engstrom, supra note 103, at 328–79. 291 Id. at 306. 292 A survey of state damages caps can be found at the American Tort Reform Association’s Web site, http://www. atra.org/legislation. Though these caps appear more often in state law, federal law incorporates a few such provisions, e.g., that found in the Oil Pollution Act of 1990, Pub. L. No. 101-380, § 1004, 104 Stat. 84, Aug. 18, 1990 (capping cer- tain forms of civil liability for oil spills). 293 CAl. Civ. Code § 3333.2 (West 1997). 294 National Childhood Vaccine Injury Act of 1986, Pub. L. No. 99-660, 100 Stat. 3755 (codified at 42 U.S.C. §§ 300aa–1 to –34 (2012)). 295 42 U.S.C. §§ 300aa–10 to –16. 296 Id. §§ 300aa–15. 297 E.g., Bruesewitz v. Wyeth LLC, 562 U.S. 223, 243, 131 S. Ct. 1068, 1082, 179 L. Ed. 2d 1, 17 (2011); Riegel v. Medtronic, Inc., 552 U.S. 312, 330, 128 S. Ct. 999, 1011, 169 L. Ed. 2d 892, 906 (2006); (2008); Geier v. American Honda Motor Co., 529 U.S. 861, 886, 120 S. Ct. 1913, 1928, 146 L. Ed. 2d 914, 935 (2000). 298 Biomaterials Access Assurance Act of 1998, Pub. L. No. 105-230, 112 Stat. 1519, 1525.

35 Communications Decency Act of 1996299 greatly limited plaintiffs’ ability to recover against the operators of Internet websites for content gener- ated by third parties. • Preemption by Regulation: Administrative regulations enacted under the authority conferred by statute also may preempt tort remedies. In Geier v. American Honda Motor Co.,300 for exam- ple, the United States Supreme Court held that air bag standards promulgated by NHTSA pre- empted a plaintiff ’s claim that her automobile was defective for failing to incorporate a driver’s- side air bag.301 The Geier Court acknowledged that the statute under which the agency had promul- gated the regulation did not expressly preempt tort remedies.302 The plaintiff ’s lawsuit was never- theless barred, the Court held, because it sought to achieve through litigation a goal (universal incorporation of air bags) that conflicted with the pertinent safety standard’s perceived objective of encouraging a variety of passive restraint systems within automobiles.303 This list does not exhaust the ways in which leg- islatures or administrative agencies can limit tort liability. Laws also may alter the procedures associ- ated with tort claims to make recovery more diffi- cult,304 provide product manufacturers with an affir- mative defense such as a statute of repose,305 or eliminate or preclude certain causes of action.306 In certain instances, however, state laws that limit plaintiffs’ remedies have been struck down as violat- ing the enacting state’s constitution.307 4. Additional Considerations Tort law does not necessarily apply to new technol- ogies in an easily predictable or wholly rational man- ner. Instead, there exist several possible sources of contingency in the application of even well-established rules to a new device. This report touched upon sev- eral of these dynamics in its discussion of innovations of the past: immature claim consciousness;308 inade- quate, skewed, or evolving risk assessments;309 absent or off-point baseline analogical references;310 idiosyn- cratic events; and path dependence in the law.311 B. Application The text below will discuss other analysts’ predic- tions regarding the civil liability prospects of driver- less vehicles, before offering a forecast of its own. 1. Other Analyses of Liability for Accidents Existing analyses of how tort liability may adhere to the manufacture and use of driverless vehicles tend to agree on certain matters. They have reached the shared conclusion that a proliferation of driver- less vehicles eventually will lead to an “upward” shift in the locus of civil liability for everyday acci- dents, away from drivers and toward the manufac- turers of these devices.312 This movement, it is believed, will necessitate greater reliance upon products liability law as a rule of decision in vehicle accident cases.313 Within this area of the law, design defect and warning defect claims are expected to be more common than manufacturing defect claims.314 299 Communications Decency Act of 1996, Pub. L. No. 104-104, § 509, 110 Stat. 133, 137–39 (Title V, Telecommu- nications Act of 1996, Pub. L. No. 104-104, 110 Stat. 56). 300 529 U.S. 861, 120 S. Ct. 1913, 146 L. Ed. 2d 914 (2000). 301 Id. at 865. 302 Id. at 868. 303 Id. at 886. 304 Colo. rev. sTAT. Ann. § 13-25-127(2) (West 2014) (permitting an award of exemplary damages only on the presentation of proof beyond a reasonable doubt that the tort was attended by fraud, malice, or willful and wanton conduct). 305 E.g., Tex. Civ. PrAC. & rem. Code Ann. § 16.012(b) (West 2015) (creating a 15-year statute of repose, running from the time of sale, in products liability actions). 306 See, e.g., Bonnie Hershberger, Supersized America: Are Lawsuits the Right Remedy?, 4 J. food l. & Pol’y 71, 82 (2008) (discussing the enactment of laws addressing “obesity lawsuits” in approximately half of the states). 307 E.g., Watts v. Lester E. Cox Medical Centers, 376 S.W.3d 633, 636 (Mo. 2012) (holding that a state law cap- ping non-economic damages in medical malpractice cases violated the right to a jury trial conferred by the Missouri state constitution). 308 See William L.F. Felstiner, Richard L. Abel & Austin Sarat, The Emergence and Transformation of Disputes: Naming, Blaming, Claiming…, 15 lAW & soC. rev. 631 (1980); Graham, Of Frightened Horses and Autonomous Vehicles: Tort Law and Its Assimilation of Innovations, supra note 92, at 1266; Michael L. Rustad & Thomas H. Koenig, Cybertorts and Legal Lag: An Empirical Analysis, 13 so. CA. inTerdisCiPlinAry L.J. 77, 77–78 (2004) (discuss- ing the concept of “legal lag” as applied to tort law and new technologies). 309 See rAChel mAines, AsbesTos And fire: TeChnologi- CAl TrAde-offs And The body AT risk (2005) (discussing the shifts that have occurred in the risk perceptions that have surrounded asbestos); Graham, Of Frightened Horses and Autonomous Vehicles: Tort Law and Its Assimilation of Innovations, supra note 92, at 1256–66; Peter Huber, Safety and the Second Best: The Hazards of Public Risk Manage- ment in the Courts, 85 ColUm. l. rev. 277, 319 (1985). 310 Graham, Of Frightened Horses and Autonomous Vehicles: Tort Law and Its Assimilation of Innovations, supra note 92, at 1252–56; Mandel, supra note 175, at 553–57. 311 See Hathaway, supra note 174. 312 RAND rePorT, supra note 181, at xxii; Marchant & Lindor, supra note 253, at 1323. 313 Julie Goodrich, Driving Miss Daisy: An Autonomous Chauffeur System, 51 hoUsTon l. rev. 265, 280–81 (2013). 314 Marchant & Lindor, supra note 253, at 1323.

36 Each of these uncertainties could have a profound effect on the civil liability rules associated with driv- erless vehicles. Greater reliance on a V2I system than is presently anticipated, for instance, likely would necessitate substantial government engage- ment in the development of the necessary infra- structure, which in turn could create a larger aper- ture for negligence claims against state and local authorities (and implication of the various immuni- ties that can apply to government decision making) than might otherwise exist. These contingencies mean that any predictions regarding the liability prospects of driverless vehi- cles must be both general and probabilistic, state key underlying assumptions, and appreciate the temporal dimension of tort law and practice. The forecast that follows assumes the gradual emer- gence, over the next several years, of driverless vehi- cles with incrementally evolving capabilities up to and including Level 4 NHTSA functionality. It fur- ther assumes that advanced driverless vehicles will be mass-marketed to consumers and eventually will garner a significant share of the market for new vehicles, but share the road with both conventional vehicles and vehicles with automated capabilities for at least the next half-century. 2. Liability for Accidents: A Staged Forecast These assumptions lead to the following plausi- ble, if not unavoidable, projections. The types of per- sonal injury cases associated with driverless vehi- cles likely will evolve over time. Claims that allege user negligence will predominate at first, but even- tually will fall off substantially as driverless vehi- cles and their users both grow more common and competent. These claims against users will be replaced, to a degree, by claims that allege defects in driverless vehicles (the “upward” shift spoken of by other commentators), although these claims will not be as common as negligence lawsuits brought against drivers are today, due to the enhanced safety profile of these devices. Most early claims against manufacturers of driv- erless vehicles likely will resemble those lodged against the makers of conventional vehicles, attack- ing matters such as a perceived lack of design “crashworthiness.” There may be a lag before a sub- stantial number of sophisticated defect claims that specifically attack driverless features and functions appear. Any early cases, however, likely will prove important in directing the future path of the law. Significantly, in the long run the total number of personal injury lawsuits involving vehicles should On other topics, existing predictions part ways. Some, although not all, of these analyses have expressed concerns that judges and juries will over- estimate the risks associated with driverless vehi- cles and fail to fully take into account the safety- enhancing characteristics of these devices.315 Of particular concern is the possibility that juries will find manufacturers of driverless vehicles liable for an aspect of a vehicle’s performance (such as a deci- sion to swerve left when confronted by a particular scenario) that is on balance preferable to the known alternatives, but nevertheless caused the particular plaintiff in the case at hand to suffer an injury. These worries have led some commentators to ask whether it is or will become good policy to preempt or limit the tort liability of the manufacturers of driverless vehicles.316 Other observers disagree with this pre- diction and prescription. Those who take the latter view anticipate that the common law will prove capable of fair application to driverless vehicles in any personal injury lawsuits that may arise, whereas the preemption of tort liability would eliminate an incentive for the manufacturers of driverless vehi- cles to improve their products’ safety.317 Presently, all of these predictions and conclu- sions—regardless of whether they agree or con- flict—seem reasonable, but hardly indisputable. It is difficult enough to forecast how the law will apply to today’s technologies. The catalogue of contingencies associated with driverless vehicles therefore allows only rough predictions. In addition to the possibility of changes in the underlying law, it is unclear whether and to what extent: • Driverless vehicles will be primarily self- directed, or will rely on V2V or V2I communica- tions to direct their movement; • The software associated with driverless vehi- cles will be marketed as, and generally understood to represent, a product distinct from a vehicle’s physical hardware; • Driverless technologies will evolve in the man- ner and sequence presently anticipated; • Driverless vehicles initially will be available only to limited audiences, or be deployed only in particular contexts (such as providing transporta- tion for hire within urban areas); and • Conspicuous accidents or other adverse events will occur that may frustrate or delay the use and acceptance of these devices. 315 Id. at 1334–35. 316 RAND rePorT, supra note 181, at 118; Goodrich, supra note 313, at 292–93. See also id. at 132 (“[M]anufac- turer liability is expected to increase, and this may lead to inefficient delays in the adoption of these technologies.”) 317 E.g., Villasenor, supra note 253. 318 See generally bosTon ConsUlTing groUP rePorT, supra note 188 (offering similar predictions regarding the diffusion of autonomous vehicles).

37 reevaluation and possible alteration of these rules― whether through federal preemption or otherwise. To the extent that litigation concerning driverless vehicles does arise during this span, the immaturity of the technology and plaintiffs’ evolving “claim con- sciousness” suggests a bias toward lawsuits that attack 1) decisions made by the drivers of automated vehicles with driverless capacities, and in particular, decisions associated with the engagement and main- tenance of driverless functionalities; 2) an alleged failure to provide sufficient warnings regarding risks associated with these devices, particularly vis- à-vis the utilization of driverless capabilities (as opposed to “hands-on” driving); and 3) alleged defects in a vehicle’s sensors, actuators, and other hardware, as opposed to defect claims that attack flaws in the software that translates information derived from sensors into driving instructions. With regard to the first of these categories of potential claims, early adopters of driverless vehicles may be subject to allegations that they failed to exer- cise the required care vis-à-vis one or more novel attributes of these devices. The first wave of auto- mated vehicles with driverless functionalities will be capable of driverless operation only in certain areas, and may be precluded by law or design from being operated in driverless mode elsewhere. This limita- tion raises the possibility of litigation in which injured plaintiffs will ascribe their injuries to an operator’s assertedly negligent decision to utilize a driverless vehicle in an area where, or at a time when, it either was not authorized for use, or when or where it may have been unreasonably unsafe to engage an auto- mated vehicle’s driverless functions.320 These claims may point to a violation of a pertinent time, place, or manner statute, regulation, or ordinance as bespeak- ing negligence, or may seek to create novel common- law “rules of the road” that will come to govern inter- actions between driverless vehicles and either conventional vehicles or pedestrians. The improper engagement of driverless features in certain areas or zones also may generate warning- defect claims against manufacturers, as well as related design-defect claims that condemn, for exam- ple, the absence of an automatic transfer of the driv- ing function to active human direction under certain circumstances. Manufacturers predictably will seek to avoid the first type of claim by requiring that pro- spective purchasers undergo extensive training and certify their awareness of various hazards and limita- tions associated with the operation of these vehicles. The second type of claim may require litigation to ascertain the manufacturer’s design responsibilities. drop precipitously, due to the ability of sophisticated driverless vehicles to avoid or reduce the severity of accidents that would befall human drivers. The text below divides this scenario into three stages, each of which signifies a different phase in the maturation of personal injury litigation involv- ing driverless vehicles. a. Stage One: Early Litigation—The immediate future likely will witness only a modest volume of tort litigation owing to the distinctive qualities of driverless vehicles. If these cases emerge, most will address basic issues associated with the use of driv- erless vehicles, such as the proper spheres allocated to human direction and automatic control of these devices, how users and manufacturers should man- age the transitions between these modalities, and core principles regarding interactions between driv- erless vehicles on the one hand and conventional automobiles and other highway users on the other. Several dynamics may contribute to a lag in cases involving driverless vehicles during this span. These constraints include the limited number of driverless vehicles on the highways; advancing but still imma- ture expectations regarding the rights and responsi- bilities of the manufacturers and operators of driv- erless vehicles (and others who come into contact with these devices); difficulties that early adopters of these devices may experience in recovering for their injuries, should a sense prevail that the devices are to some degree still experimental and choices to use them assume significant risk; statutes and regu- lations that restrict the use of driverless vehicles and technologies incorporated within these vehicles; and potential marketing and sales strategies associ- ated with the first wave of driverless vehicles, such as a practice of channeling sales toward institu- tional customers who stipulate to operate these devices only in a manner whereby accidents are par- ticularly unlikely to occur.319 The rarity of cases involving driverless vehicles, however, will make any cases that are litigated espe- cially significant; indeed, perhaps unrealistically important. During this early phase, due to path dependence and the likely notoriety of initial judi- cial decisions that concern this new technology, whatever tort litigation does occur may have sub- stantial impact on the development of the law per- taining to these vehicles. For example, if early case outcomes suggest that generic tort principles do not properly account for the unique risks and benefits of driverless vehicles, pressure will build for the 319 See Bryant Walker Smith, Proximity-Driven Liabil- ity, 102 geo. L.J. 1777, 1815–18 (2014) (considering how entities that sell driverless vehicles or services associated with these devices could manage their risk profile through alternative sales and service strategies). 320 See RAND rePorT, supra note 181, at 132 (discussing the “weak spot” at the driver-vehicle interface).

38 especially in jurisdictions that permit the introduc- tion of such remedial measures in court.324 As suggested before, depending upon their num- ber, cost, and outcome, early design and warning defect cases may cause manufacturers to press for liability-lessening measures. One such option would involve the creation of “safe harbors” through legislation or regulation, whereby the sat- isfaction of a safety standard would provide an affirmative defense to liability. Alternatively, legis- latures could preempt state tort liability standards for driverless vehicles—either altogether, as to a particular type of defendant, or for particular types of accidents. Whether and to what extent these efforts will prove successful will depend upon a constellation of factors, including the per- ceived safety benefits of driverless vehicles, the perception that liability may or may not unduly deter their development and deployment, and the persuasiveness of affected stakeholders. In summary, the first phase of litigation over driver- less vehicles is expected to involve relatively few cases. Most litigation against the makers of driver- less vehicles during this span probably will not con- centrate upon driverless capabilities at all, but instead will seem almost identical to claims pres- ently lodged against the makers of conventional automobiles. As for claims against the users of driv- erless devices, plaintiffs may try to restyle existing theories of negligence commonly directed against the drivers of conventional vehicles, for example, by attacking the users of driverless vehicles as insuffi- ciently attentive toward circumstances that argu- ably required “hands-on” driving. This sort of claim may lead to the integration of manufacturers into the litigation mix, under a failure-to-warn theory. It also is possible that even in this early phase of liti- gation, manufacturers may become enmeshed in cases that challenge specific design choices made regarding technologies such as sensors, and, con- ceivably, programming decisions regarding how the vehicle should respond to stimuli. Depending upon their number and cost, these lawsuits may cause manufacturers to press for laws or regulations that will provide affirmative defenses to liability, or out- right preemption of state tort liability. b. Stage Two: The Maturation of Driverless-Vehicle Litigation.—Operating mostly within the basic framework of rules produced by initial litigation, the types of civil liability claims associated with Other design-defect claims also may emerge. While basic driverless technologies are still devel- oping, the risk-utility profiles associated with alter- native design choices may be difficult to pinpoint. Some design decisions made by manufacturers may be susceptible to scrutiny sooner than others, however. Sensor technology, for example, is (or soon will be) at a point where different designs can be meaningfully compared with one another, as would be required for recovery for a design defect in risk- utility jurisdictions.321 Difficulties in intelligently critiquing program- ming choices may inhibit nuanced design-defect lawsuits involving vehicle software, at least for a time. Software may produce early and easy product- defect litigation where it leads to palpably improvi- dent outcomes—such as a vehicle turning abruptly and unexpectedly into oncoming traffic, running a red light, or crossing over a sidewalk when making a turn—in which case a defect of some sort will be difficult to deny.322 In such cases, the presence of a defect likely will be ascertained simply by assessing whether a vehicle’s actions substantially deviated from customary safe driving practices utilized by the closest substitute—human drivers under simi- lar conditions. That said, certain features of driverless vehicles may accelerate the normally gradual process through which plaintiffs develop claim conscious- ness and their counsel develop the ability to iden- tify and attack a design defect. Highway users already are conditioned to regard an automobile accident as the potential basis for a “claim” of some sort.323 Furthermore, information collected or relayed by driverless vehicles may shorten the feedback loop through which data regarding a product’s dangerous qualities is gathered and translated into possible improvements. Even the issuance of a software “patch” for a driverless vehi- cle could greatly simplify a plaintiff ’s lawyer’s bur- den of locating and proving a product defect, 321 bosTon ConsUlTing groUP rePorT, supra note 188, at 13. 322 Per the resTATemenT (Third) of TorTs: ProdUCTs liAbiliTy supra note 269: It may be inferred that the harm sustained by the plaintiff was caused by a product defect existing at the time of sale or distribution, without proof of a specific defect, when the incident that harmed the plaintiff: (a) was of a kind that ordinarily occurs as a result of product defect; and (b) was not, in the particular case, solely the result of causes other than product defect existing at the time of sale or distribution. Id., § 3. 323 hensler, supra note 255, at 122–23. 324 Compare fed. r. evid. 407 (barring the introduction of evidence regarding a subsequent remedial measure to prove the existence of product defect) with Ault v. Int’l Harvester Co., 13 Cal. 3d 113, 120, 528 P. 2d 1148, 1152, 117 Cal. Rptr. 812, 816 (1974) (permitting the introduction of this evidence).

39 these sorts of decisions, and have their conduct reviewed after the fact for reasonableness by judges and juries. Yet the need to have these matters resolved in advance within a vehicle’s software code may provide a basis for manufacturers or software suppliers to push for the promulgation of liability- limiting standards for the algorithms used in these contexts. If adopted, these rules may lead to addi- tional safety standards involving other coding mat- ters. Alternatively or in addition, judges may have substantial gatekeeping roles assigned to them in determining whether alternative programming decisions that would address these scenarios repre- sent “reasonable” alternative designs, as many juris- dictions require for presentation of a design-defect claim to a jury.326 Other claims that may appear within this second phase will be even more novel, and are therefore more difficult to anticipate. Some of these claims may involve intangible harms. The prodigious amount of data generated by driverless vehicles, together with the connected attributes of these devices, will create incentives and opportunities for businesses and indi- viduals to collect and use this information for profit, or for other purposes. Consumers or other affected persons may regard this exploitation of “their” data as injurious. Although most of the resulting cases will be resolved by reference to evolving consumer protec- tion and privacy statutes, the common law of torts, and specifically the privacy torts,327 may be invoked as a means of deciding who can properly control this information. Some of these disputes may challenge the collection or dissemination of information by vehicle manufacturers or software providers. Other claims may involve “hacking” by hostile third par- ties.328 And still other potentially tortious fact pat- terns, yet unknown, may arise. Litigation during this period also may chal- lenge distinctions that historically have appeared within products liability law. For example, while products are subject to strict products liability, driverless vehicles likely will evolve and mature dur- ing a second phase of litigation, in sync with the increasing capabilities and prevalence of these devices. Some of the tort claims brought within this second stage of litigation will resemble those pur- sued in earlier cases. Operators of driverless vehicles likely will continue to face claims that they improp- erly engaged or utilized a vehicle’s driverless func- tions, and manufacturers will have to respond to charges that their products’ warnings or designs facilitated these errors. This period also may witness the emergence of more sophisticated claims against users, such as allegations that users overrode safety directions programmed within the vehicle or selected an unreasonably aggressive driving mode. Meanwhile, as driverless vehicles grow increas- ingly sophisticated and take over ever more respon- sibility relative to their human occupants, the prin- cipal locus of liability for accidents is expected to transition away from people using these vehicles for transportation and toward the manufacturers of these devices and the software used in them. As this shift occurs, products liability cases will appear that spot and attack increasingly subtle defects in the software and hardware that direct a vehicle’s move- ments and actions. The evolution of software, the development of expectations regarding how it should perform, and the (possible) availability of data that allows for comparison across software platforms will enhance the ability of plaintiffs to pursue design- defect claims in which a particular programming choice associated with an accident is attacked as defective.325 Especially during the early portion of this phase, some of these claims may concern the combined operation of sensors (what the vehicles should have observed) and software (how the vehicle should have responded to the data it received). As time passes, the emphasis placed upon the latter portion of this equation likely will increase, as plain- tiffs and their counsel grow more comfortable in challenging software design decisions. Once it reaches a sufficient stage of sophistica- tion, litigation over the defectiveness of vehicle soft- ware may present difficult technical and moral issues to judges and juries. Programmers of driver- less vehicles will have to decide in advance how the vehicle will respond to certain situations in which some sort of accident is unavoidable. In these sce- narios, a particular coding decision may reduce the risk of harm to the driver, but impose greater risk upon a passenger or third parties, such as pedestri- ans (or vice versa). Human drivers already make 326 resTATemenT (Third) of TorTs: ProdUCTs liAbiliTy, supra note 269, at § 2 cmt. d. 327 See William L. Prosser, Privacy, 48 CAl. l. rev. 383 (1960) (recognizing four distinct common-law privacy torts: intrusion, public disclosure of private facts, false light in the public eye, and (mis)appropriation of one’s likeness). 328 See Brian Leon, Students from China’s Zhejiang Uni- versity Successfully Hack a Tesla Model S to Win $10,000, N.Y. dAily neWs (Aug. 8, 2014), http://www.nydaily news.com/autos/chinese-university-students-successfully- hack-tesla-model-s-article-1.1896540 (discussing how, in response to a challenge and prize offer, within hours a team of university students used a computer to “control[] the [Tesla’s] lights, horn, sunroof, and door locks remotely all while the car was in motion”). 325 David C. Vladeck, Machines Without Principals: Liability Rules and Artificial Intelligence, 89 WAsh. l. rev. 117, 132 (2014).

40 personal injury claims involving vehicles, owing to the safety benefits of driverless vehicles relative to conventional vehicles and the former’s replacement of the latter on the nation’s highways. It is unclear whether persons who suffer injuries associated with driverless vehicles during this mature stage will seek relief primarily through the courts, or through other avenues. As just noted, the anticipated decline in the frequency and severity of vehicle accidents will be accompanied by a prolif- eration of parties or entities potentially contribut- ing to the remaining accidents (OEMs, program- mers, hardware suppliers, state and federal municipalities, and providers of apps and V2V and V2I communications). If these groups are routinely added to the litigation mix and thereby complicate the liability equation, a different compensation system may well recommend itself. Although no- fault insurance has presented its own set of prob- lems, a form of no-fault insurance may better fit an age of driverless vehicles than it does the present fault- and defect-based legal regime. The National Childhood Vaccine Injury Act of 1986 offers yet another model for an alternative compensation system that may emerge. 3. Conclusions The common grounds for civil liability for per- sonal injuries associated with the manufacture, use, and operation of driverless vehicles likely will evolve over time. Early lawsuits will draw heavily from existing law that relates the rights and responsibili- ties of the makers and users of conventional vehi- cles. Claims gradually will grow more sophisticated and begin to critically evaluate the capabilities of driverless vehicles as a distinct technology. As other commentators have noted, as primary responsibility for decision making while driving shifts from human drivers to driverless vehicles, the principal repository of liability for everyday traffic accidents correspondingly will drift away from indi- vidual vehicle operators and toward product manu- facturers. Automobile-accident plaintiffs in the future presumably will rely increasingly on the strict-liability theories of recovery that are available against defendants involved in the supply chain for products, instead of the negligence principles that apply to human drivers. Over time, driverless vehi- cles may lead to changes in generic products liability doctrine, although the precise direction of these adjustments is difficult to anticipate. Eventually, driverless vehicles likely will result in a significant reduction in the total number of lawsuits involving the operation of motor vehicles. Negligence claims against users may remain some- what more prevalent and persistent than what services are not.329 Under prevailing case law, aeronautical charts represent a product,330 while chauffeurs are regarded as providing a service. Given this divide, it is unclear whether and when the software incorporated within driverless vehi- cles will be regarded as a product or as a service.331 Also, the post-sale responsibilities of product manufacturers are governed mostly by the law of negligence. In applying this standard, courts have not imposed significant responsibilities upon manufacturers to update products that were not defective at the time they were sold. This too may change, since yesterday’s programming decisions for driverless vehicles may produce unreasonable dangers within a very short period of time, and it is expected that software updates for these vehi- cles will be capable of delivery almost instanta- neously and at low cost.332 Finally, depending on the path that driverless vehicle technologies follow, during this time new claims may emerge against third parties who nei- ther use nor manufacture driverless vehicles, but are in other ways responsible for their operation on the highways. If a V2I connected vehicle infrastruc- ture becomes prevalent, local governments may face defective programming claims. Likewise, there may appear new business niches associated with the operation of driverless vehicles, such as coun- terparts to present-day navigation “apps,” that promise to enhance the efficiency or otherwise direct the performance of driverless vehicles manu- factured by a different company. Original manufac- turers may argue that use of these “apps” amounts to the misuse or alteration of a product, which may provide them with a defense to a tort action. The makers and retailers of these operation enhancers, meanwhile, may become the subjects of litigation and regulation. c. Stage Three: A Mature Claiming Environ- ment.—At some point, personal injury litigation associated with driverless vehicles likely will reach a mature state. Novel issues still may appear, but not as often, and most tort claims involving driver- less vehicles will become routinized, as occurred between the 1920s and the 1960s with claims involv- ing conventional automobiles.333 This routinization will affect a diminished and ever-shrinking pool of 329 resTATemenT (Third) of TorTs: ProdUCTs liAbiliTy, supra note 269, § 19. 330 See Abney, supra note 286. 331 See Smith, supra note 319, at 1817–19. 332 See id. at 1803–08. 333 Samuel Issacharoff & John Fabian Witt, The Inevi- tability of Aggregate Settlement: An Institutional Account of American Tort Law, 57 vAnd. l. rev. 1571, 1618 (2004).