Liability from the use of medical artificial intelligence: a comparative study of English and Taiwanese tort laws

Background: Modern artificial intelligence applications are appearing in healthcare and medical practices. Artificial intelligence is used both in medical research and on patients via medical devices. The aim of this paper is to examine and compare English and Taiwanese tort laws in relation to medical artificial intelligence. Methods: The methodologies employed are legal doctrinal analysis and comparative law analysis. Results: The investigation finds that English tort law treats wrong diagnostic or wrong advice as negligent misstatement, and mishaps due to devices as a physical tort under the negligence rule. Negligent misstatement may occur in diagnosis or advisory systems, while a negligent act may occur in products used in the treatment of the patient. Product liability under English common law applies the same rule as negligence. In Taiwan, the general principles of tort law in Taiwan’s Civil Code for misstatement and negligent action apply, whereas the Consumer Protection Act provides for additional rules on product liability of traders. Conclusions: Safety regulations may be a suitable alternative to tort liability as a means to ensure the safety of medical artificial intelligence systems.

Two Advantages of the Negligence Rule Over Strict Liability when the Parties are Risk Averse

When parties are risk-averse and therefore take out insurance, the efficiency of a tort rule depends on how well the insurance contracts govern incentives, risk allocation and transaction costs under the rule. This article presents two overlooked or discarded advantages of the rule of negligence over strict liability, which appear when insurance contracts are incomplete due to ex-ante transaction or ex-post verification costs. One advantage arises because of a legal impediment under strict liability: insurers cannot exempt coverage for all acts of simple negligence. Instead, the insurer must, at a cost, precisely specify each act for which coverage is excluded. Such specification can be prohibitively costly when there are many acts and many contingencies. These transaction costs, or the inefficient risk allocation associated with a deductible, are avoided under the negligence rule, where under idealized conditions the injurer can simply take due care and need not take out insurance. The other advantage of the negligence rule is that it provides incentives for the victim to bring forward information about the injurer’s acts. The victim has little incentive to convey such information under strict liability, whereas the victim’s insurer may elicit it, e. g. by not covering the victim’s loss fully.

The Mistaken Restriction of Strict Liability to Uncommon Activities

Courts generally insist that two criteria be met before imposing strict liability rather than basing liability on the negligence rule. The first—that the injurer’s activity must be dangerous—is sensible because strict liability possesses general advantages over the negligence rule in controlling risk. But the second—that the activity must be uncommon—is ill-advised because it exempts all common activities from strict liability no matter how dangerous they are. Thus, the harm generated by the large swath of common dangerous activities—from hunting, to construction, to the transmission of natural gas—is inadequately regulated by tort law. After developing this theme and criticizing ostensible justifications for the uncommon activity requirement, the article addresses the question of how it arose. The answer is that its legal pedigree is problematic: it appears to have been invented by the authors of the first Restatement of Torts. The conclusion is that the uncommon activity requirement for the imposition of strict liability should be eliminated.

Eliciting Multi-Dimensional Care through Liability and Regulation

This paper examines the incentive performance of liability and regulation when a potential injurer can take two types of preventative care, one of which is observed and one of which is not. The problem is studied in a general asymmetric information model, where settlement behavior is endogenous and which incorporates an uncertain legal standard. Contrary to existing literature, we find (1) a shift to a negligence rule may have a perverse effect on unobserved care; (2) uncertainty in legal standards may be socially beneficial as it provides a relatively good incentive for unobserved care; (3) a regulation may not be effective if preventative care efforts are substitutes but is effective if they are complements; (4) an increase in settlement rate may or may not increase the level of care, depending on the cause of the increase; (5) a “decoupling” arrangement with a feature that the defendant pays more than the plaintiff recovers, reduces legal costs and is therefore socially beneficial.

Uncertainty Regarding Interpretation of the “Negligence Rule” and Its Implications for the Efficiency of Outcomes

There are two ways that the negligence rule is interpreted. Under one interpretation a negligent injurer is liable for the entire harm to the victim; and under the other interpretation a negligent injurer is liable only for that part of the harm which can be ascribed to his negligence. Both these versions are efficient. However, if there is uncertainty regarding whether the court will be employing the full liability version or the incremental liability version for determining the liability of a negligent injurer, notwithstanding the fact that both the versions are efficient, inefficiency is possible. It is shown in the paper that a necessary and sufficient condition for efficiency in all cases is that the subjective probability with which the injurer expects the standard version to be employed must be greater than or equal to the subjective probability with which the victim expects the standard version to be employed. For the subset of applications without complementarities in the cares of the two parties and which are such that the total social costs are minimized at a unique care-configuration, it is shown that efficiency obtains regardless of the subjective probabilities with which the parties expect the two versions. One very important conclusion that emerges from the analysis of this paper is that when courts employ more than one liability rule, even if all the employed rules are efficient, the efficiency of all outcomes cannot be taken for granted merely on the ground of the efficiency of the employed rules.

Risk Aversion, the Hand Rule, and Comparison between Strict Liability and the Negligence Rule

One of the most important propositions in the economics of liability rules states that strict liability and the negligence rule are equivalent and first-best efficient if the standard of due care is set according to the Hand rule. This proposition hinges on the assumption that individuals are risk neutral. This paper considers this proposition with risk-averse individuals, and demonstrates that the proposition does not extend. In particular, the two liability rules are not equivalent, and the analysis compares the two liability rules in terms of utilitarian social welfare.

What Mistake Of Law Just Might Be

This Article examines and evaluates a distinctive, increasingly popular account of the Mistake of Law doctrine. The doctrine, deeply ingrained in American criminal law, is at the same time notoriously unclear in its scope, content, and application. A growing number of legal theorists have criticized the traditional interpretation of the doctrine; legal moralists in particular have argued that this account is conceptually confused. Because the doctrine’s use of a strict liability punishment regime does not incentivize individuals to learn the law as well as a negligence regime might, legal moralists argue that the doctrine cannot be explained by a desire to incentivize legal knowledge. In evaluating this argument, the Article defends the traditional account, often identified with the liberal positivism of Justice Holmes. The Article advances three main arguments. First, legal moralism’s claim that a negligence standard is more effective than a strict liability standard in incentivizing individuals to learn the law is false: the safe harbor provision of a negligence rule acts as an insurance effect, disincentivizing individuals to learn the law. Second, legal moralism assumes that the moral content of the criminal law is determinate, and that agents have perfectly rational, objective motivational sets. These are illicit assumptions that result in a flawed argument. Finally, the Article contends that legal moralism misinterprets the structural core of the traditional account: properly understood, the Mistake of Law doctrine employs a negligence–strict liability hybrid, and is thus more sophisticated than legal moralists realize. The Article concludes that, contrary to what a surprising number of criminal law theorists have come to accept, legal moralism fails to make a plausible case against the traditional account of the Mistake of Law doctrine.

Lapses and Substitution

This chapter deals with lapses of attention, for example by a driver or a doctor, and explains how a negligence rule gives injurers an incentive to substitute activities with unavoidable accidents for activities with lapses. Under current tort law, a lapse of attention will always be considered negligence and trigger liability for the resulting harm. However, the chapter shows that it should not be the case. After providing an overview of lapse defenses in prevailing law, the chapter examines the openness of liability law to the lapse defense and some activities that substitute unavoidable harm for lapses. It also explains how a lapse defense is implemented and shows how the lapse defense dampens inefficient substitution and increases collection of supporting information.