Book Read Free

The Economics of Artificial Intelligence

Page 83

by Ajay Agrawal


  i

  i

  incorporating technology i into the fi rm’s product, and r is the probability i

  2. It is important to note that this chapter focuses solely on the likely impacts on innovation and the direction of technological change. We refer interested readers to Hay and Spier (2005) and Polinsky and Shavell (2010) for an overall welfare discussion of the liability system and its features, and to Marchant and Lindor (2012) and Hubbard (2015) and the references within for details of tort law and an exploration of their applications to autonomous vehicles and sophisticated robots.

  Economics of Tort Liability and Innovation in Artifi cial Intelligence 495

  that the use of the product will result in personal injuries. The expected lia-

  bility cost given that injury happens is H , which captures the (conditional)

  probability that a liability suit will be fi led and the expected cost that the

  fi rm will face if involved in such a suit. We expect H to be positive even if the fi rm is fully insured against claims for monetary damages because liability suits also invariably result in opportunity costs of employee time and

  fi rm resources, as well as in reputational damage.

  The fi rm’s expected profi t, net of liability risk from selling a product incor-

  porating technology i, is

  = b r H.

  i

  i

  i

  We denote the technology that the fi rm currently uses as O and consider

  the fi rm’s decision to develop a new technology, which we denote as N. We

  assume a simple R&D process such that successful development takes place

  with probability p(x) = x if the innovator incurs a research cost C(x) = x 2 / 2.

  As in Aghion et al. (2016), we refer to x as the “innovation intensity,” which

  captures the likelihood of successfully developing a new technology. In this

  setting, the problem for the innovating fi rm is

  x 2

  max x

  + (1 x)

  N

  O

  x

  2

  which yields the following:

  (1)

  x* =

  = b

  b + ( r

  r ) H.

  N

  O

  N

  O

  O

  N

  Formula (1) provides some basic insights into the relationship between

  liability and innovation. First, at the intensive margin, the sign of the deriva-

  tive of x* with respect to H captures the directional eff ect of an increase in liability risk on innovation intensity. Thus, an increase in liability risk

  suppresses innovation incentives for new technologies that are riskier than

  the current technology ( r > r ) but encourages new technologies that are N

  O

  safer ( r < r ). In other words, changes in liability risk aff ect the type of N

  O

  technologies in which a fi rm invests and infl uence the direction of innova-

  tion. Second, investment in innovation takes place if the profi t potential of

  the new technology is greater than its liability risk, relative to the old tech-

  nology—that is, x* > 0 only if b – b > ( r – r ) H. Thus, at the extensive N

  O

  N

  O

  margin, marginal changes in liability risk will not aff ect whether the fi rm

  develops the new technology if it is expected to be highly profi table relative

  to the existing one (i.e., b – b is very large) unless it is extremely risky. In N

  O

  contrast, liability concerns will matter more for technologies “at the margin”

  (i.e., the improvement in expected profi tability is modest).

  Galasso and Luo (2017) extend this stylized model to the medical setting,

  in which physicians (i.e., the direct users of technologies) face malpractice

  liability risk. Changes in their liability exposure aff ect innovation incentives

  in medical technologies through the demand channel. Assuming that ideas

  496 Alberto Galasso and Hong Luo

  for new technologies ( b , r ) are random draws from a bivariate distribution N

  N

  (as in Scotchmer 1999), they show that the overall eff ect of tort reforms that

  reduce physicians’ liability risk on innovation incentives is ambiguous and

  depends on the characteristics of the existing technology ( b – r ).

  N

  O

  The main message of this illustrative model is that the link between lia-

  bility and innovation is more complex and nuanced than the simple view of

  “liability chills innovation,” which ignores the potential encouraging eff ect

  of liability risk on a potentially broad set of innovations that help fi rms and

  their customers manage risk.

  20.3 Empirical Evidence on Liability and Innovation

  In a pioneering study, Viscusi and Moore (1993) examine the relationship

  between product- liability insurance costs and fi rms’ research and develop-

  ment (R&D) investments, using a data set covering large US manufacturing

  fi rms in multiple industries between 1980 and 1984. They document a sig-

  nifi cant positive correlation between the expected liability insurance costs

  and fi rms’ R&D intensity when such costs are low or moderate. Only when

  liability costs are very high the correlation is negative. Furthermore, the

  liability- innovation link is driven mainly by product rather than by process

  R&D. They interpret these results as evidence that, on average, product

  liability, rather than discouraging innovation, promotes fi rm investment in

  product safety (likely through product design).

  Galasso and Luo (2017) examine whether tort reforms that reduce physi-

  cians’ liability exposure to medical malpractice litigation aff ect incentives

  to develop new medical technologies. Diff erent from the focus on product

  liability in Viscusi and Moore (2017), they examine how liability costs that

  users (physicians) face aff ect upstream research investment. It is worth not-

  ing that such a perspective broadens the scope of innovation from product

  safety design to include a wide variety of complementary technologies that

  help physicians manage risk, such as monitoring and diagnostic devices

  and devices used in complex procedures to reduce the likelihood of adverse

  events. Because these technologies are not themselves subject to product lia-

  bility claims, they are more likely to be infl uenced by changes in user liability

  through the demand channel than by product liability.

  Using a panel data set for the period of 1985– 2005, Galasso and Luo

  (2017) fi nd that, on average, the introduction of noneconomic damage caps

  in a state is associated with a 15 percent reduction in medical device pat-

  enting. The eff ect is, however, highly heterogeneous: tort reforms have the

  largest negative impact in medical fi elds in which the probability of a mal-

  practice claim is the highest, and they do not seem to aff ect patenting of the

  highest or the lowest quality. These results are consistent with the idea that

  the decline in innovation is driven primarily by the reduced demand from

  Economics of Tort Liability and Innovation in Artifi cial Intelligence 497

  physicians for safer technologies or complementary technologies that help

  them manage risk. The welfare loss from such a large decline in quantity,

  however, appears not as worrying because patents wi
th the highest impacts

  are not negatively aff ected.

  Galasso and Luo (2018) study the medical implant industry in the early

  1990s, during which the liability risk faced by raw material suppliers signifi -

  cantly increased relative to the risk faced by downstream producers. Vitek

  was a leading producer of jaw (temporomandibular joint) implants in the

  1980s. Its Food and Drug Administration (FDA)- approved products were

  considered state of the art and safe for use by oral surgeons across the

  United States (Schmucki 1999). In the late 1980s, unexpected and wide-

  spread problems arose with Vitek’s products. Vitek fi led for bankruptcy

  in 1990 under a deluge of lawsuits. Following Vitek’s bankruptcy, implant

  patients started to fi le a large number of lawsuits against DuPont, a raw

  material supplier for Vitek’s implants and a large fi rm with “deep pockets.”3

  The consensus among industry observers is that these events generated

  a substantial increase in the perceived liability risk faced by fi rms that sup-

  plied materials to producers of permanent implants, many of which had

  withdrawn from this market. This view is well summarized in a 1994 report

  on the status of the biomaterial market (Aronoff 1995), which links this fear

  of product liability suits to the jaw implant litigation. Eventually, DuPont

  won all the lawsuits, but the process took ten years and cost over $40 million

  (House of Representatives 1997). In contrast, DuPont’s revenue from these

  implants totaled only a few thousand dollars.

  Galasso and Luo (2018) compare the rate of patenting in implant

  devices—excluding technologies involved in these litigations—to patenting

  in a control group of nonimplant medical technologies whose suppliers were

  not aff ected by the heightened litigation risk. The diff erence- in-diff erences

  (DID) results show, overall, a substantial decrease in the number of new

  patents for implants in the fi ve years after Vitek’s bankruptcy in 1990. Time-

  specifi c eff ects show that implant and nonimplant technologies exhibited

  parallel increasing trends before 1990, and that the negative eff ect on implant

  technologies was immediate after 1990 and increased in magnitude over

  time. The signifi cant drop in innovation in medical implants appears to have

  been largely driven by device producers’ expectation of a supply shortage

  of material inputs.

  To address this problem, in 1998 the US Congress passed the Biomaterial

  Access Assurance Act (BAAA), which exempts biomaterial suppliers for

  medical implants from liabilities as long as they do not engage in the design,

  3. In parallel, problems also surfaced with silicone breast implants. Also in this case, a leading implant manufacturer fi led for bankruptcy, and silicone suppliers were named as defendants in numerous lawsuits (Feder 1994).

  498 Alberto Galasso and Hong Luo

  production, testing, and distribution of the implants. The BAAA is one of

  the few federal liability reforms, an area of legislation typically reserved for

  the states (Kerouac 2001).4

  Together, the empirical evidence in Viscusi and Moore (1993) and Galasso

  and Luo (2017, 2018) challenges the simple view that “liability chills innova-

  tion.” All three papers suggest that the link between liability and innovation

  depends on the context, including the nature of the innovation, the level

  of the liability risk, and the value of the technology. Furthermore, liability

  risk aff ecting one area may impact innovation incentives in other, vertically

  related segments. More research is needed to understand the complex and

  nuanced links between liability and innovation, and whether targeted poli-

  cies can address these issues.

  20.4 Tort Liability and the Development of AI Technologies

  The liability system may aff ect innovation incentives of AI technologies

  and sophisticated robots in multiple ways, and the development of these

  technologies may, in turn, demand adjustments to the law. Below, we focus

  on a number of areas and highlight some of the economic trade- off s that

  deserve further examination, both theoretically and empirically.5

  20.4.1 Allocation of Liability Risk between Producers and Consumers

  A central question in designing a liability system for AI technologies is

  how liability risk should be allocated between producers and consumers,

  and how this allocation might aff ect innovation. Eff ective policies would

  require a basic understanding of the relationship between humans and AI

  technologies—for example, whether they are substitutes or complements

  (Agrawal, Gans, and Goldfarb, chapter 3, this volume ).

  A key promise of AI technologies is to achieve autonomy. With less room

  for consumers to take precautions, the relative liability burden is likely to

  shift toward producers, especially in situations in which producers are in

  a better position than individual users to control risk. For example, the

  operator of a fl eet of self- driving cars would have the data and predictive

  capability to provide instantaneous warnings of an adverse event. The cost

  4. Examples of such federal policies include the General Aviation Revitalization Act of 1994, which exempts makers of small aircraft from liability for planes after eighteen years, and the National Childhood Vaccine Injury Act of 1986, which limits liability for drug companies and creates a no- fault compensation system for those injured by vaccines.

  5. It is important to note that the likely impacts on innovation incentives would also depend on fi rms’ ability to write contracts and the development of the insurance markets (Schwartz 1988). We leave the discussion of these important topics and the interplay between liability law and contract law for future work. In situations in which externality (harm to third parties) is high and in the early stage of AI technologies, during which the insurance market may not be well developed or even exist, the roles of these systems are likely to be more limited than for mature technologies.

  Economics of Tort Liability and Innovation in Artifi cial Intelligence 499

  of observing systematic, hazardous user behaviors may also become suffi

  -

  ciently low such that it would be more effi

  cient for producers to take precau-

  tions through product redesign. How such a shift might aff ect innovation

  incentives would depend on how producer liability is specifi ed, especially

  whether the long- term social benefi ts are included in the analysis of the

  producer’s liability.

  On the other hand, during the transitional period of an AI technology,

  substantial human supervision may still be required. Such interaction be-

  tween AI and humans may not be obvious and diffi

  cult to predict. For ex-

  ample, it may actually be more diffi

  cult for drivers to sustain a safe degree of

  concentration levels and reaction speed when they are not actively engaged

  in driving.6 Human- machine interactions may also become more extensive

  and span increasingly complex domains as technologies are developed to

  enhance human skills. In the case of robot- assisted surgeries, for example,

  physicians may not have enough incentive to obtain suffi

  cient training or

  to be suffi

  ciently prepared for back-up options if the machine were
to mal-

  function.

  In many of these situations, it may be impractical or too costly for pro-

  ducers to monitor individual users and to intervene. Therefore, it would be

  important to maintain consumer liability to the extent that users of AI tech-

  nologies have suffi

  cient incentives to take precautions and invest in training,

  thus internalizing potential harm to others. When negative externalities are

  suffi

  ciently high, regulators may fi nd it necessary to mandate such invest-

  ment. For example, a special driver’s license may be required to operate

  a self- driving car. Similarly, doctors may be required to take a minimum

  number of training sessions with the robotic system before being allowed

  to perform certain types of procedures on patients.7

  Consumer liability may incentivize users themselves to innovate in ways

  that help them take more eff ective precautions (Von Hippel 2005). For ex-

  ample, hospitals may redesign the operating room process or reorganize phy-

  sicians’ training and work schedules. Furthermore, consumer liability may

  also incentivize producer innovation because users would demand safer and

  easier- to-use design features (Hay and Spier 2005), and mandatory train-

  ing would favor “easier to teach” designs in order to reduce adoption costs.

  20.4.2 Federal

  Regulation

  Another key issue is whether Congress should pass federal regulations on

  the safety of AI and robotic technologies that preempts state laws and how

  such regulation would aff ect innovation. This would involve the creation of

  6. The National Transportation Safety Board determined that the 2016 fatal Tesla crash was partly due to the driver’s inattention and overreliance on vehicle automation despite manufacturer safety warnings.

  7. Some expert robotic surgeons and many surgical societies have voiced the need for basic, standardized training and certifi cation in robotic surgery skills (O’Reilly 2014).

  500 Alberto Galasso and Hong Luo

  a centralized regulatory system similar to the FDA for drugs and high- risk

  medical devices: federal regulatory bodies would specify the safety stan-

  dards, and approved products would be exempted from state liability claims

  under certain conditions.8 For autonomous vehicles, the House passed a

 

‹ Prev