Civil Liability for Artificial Intelligence
The Principles project will focus on the core problem of physical harms (bodily injury and property damage). Other types of harm, such as copyright infringement, defamation, and privacy, have their own distinctive doctrinal questions and are the subjects of separate, ongoing Restatement projects. By focusing on physical harms, the project can maintain a clear scope and avoid overlap with other ongoing work. As the project progresses, the Institute will consider the broader implications of AI-caused harms and whether a more comprehensive approach might be necessary in the future.
From the press release announcing the project in 2024:
“Artificial intelligence has become front-page news, and in a short time has seen rapid advancements and increasing integration in many aspects of our society,” said ALI Director Diane P. Wood. “As AI systems become more sophisticated and capable, legal questions surrounding their use, including exposure to liability and ethical implications, are becoming increasingly complex and pressing. Given the anticipated increase in AI adoption by many industries over the next decade, now is an opportune time for The American Law Institute to undertake a more sustained analysis of common-law AI liability topics through a Principles project.”
“Courts are already facing the first set of cases alleging harms, largely related to copyright and privacy, stemming from chatbots and other generative AI models,” added Reporter Geistfeld, “ but, there is not yet a sufficient body of caselaw that could be usefully restated. Meanwhile, influential state legislatures are actively considering bills addressing AI, and Congress and federal regulators pursuant to President Biden’s Executive Order 14110 are also addressing these matters. These efforts could benefit from a set of principles, grounded in the common law, for assigning responsibility and resolving associated questions such as the reasonably safe performance of AI systems.”
“This project can help courts, the tech industry, and federal regulators understand the legal implications of AI,” explained Wood. “It focuses on common-law principles of responsibility, which can guide decision-making in the absence of applicable legislation. By identifying these principles, the project can help avoid conflicts between federal and state laws and provide clarity for all involved parties.”
The Principles project will focus on the core problem of physical harms (bodily injury and property damage). Other types of harm, such as copyright infringement, defamation, and privacy, have their own distinctive doctrinal questions and are the subjects of separate, ongoing Restatement projects. By focusing on physical harms, the project can maintain a clear scope and avoid overlap with other ongoing work. As the project progresses, the Institute will consider the broader implications of AI-caused harms and whether a more comprehensive approach might be necessary in the future.
“There are certain characteristics of AI systems that will likely raise hard questions when existing liability doctrines are applied to AI-caused harms,” explained Geistfeld. “Examples include the general-purpose nature of many AI systems, the often opaque, ‘black box,’ decision-making processes of AI technologies, the allocation of responsibility along the multi-layered supply chain for AI systems, the widespread use of open-source code for foundation models, the increasing autonomy of AI systems, and their anticipated deployment across a wide range of industries for a wide range of uses.”
Reporters
Mark Geistfeld
Mark Geistfeld, Reporter, Civil Liability for Artificial Intelligence
Mark Geistfeld is the Sheila Lubetsky Birnbaum Professor of Civil Litigation at New York University School of Law, where his research has extensively addressed the common-law rules governing the prevention of and compensation for physical harms. He has authored or co-authored five books along with over 50 articles and book chapters, often showing how difficult doctrinal issues can be resolved by systematic reliance on the underlying legal principles.
Ketan Ramakrishnan
Reporter, Civil Liability for Artificial Intelligence
Ketan Ramakrishnan is an associate professor of law at Yale Law School. His interests include torts, AI regulation, constitutional law, contracts, property, and moral and legal philosophy. His articles on these subjects are published or forthcoming in the Harvard Law Review, University of Chicago Law Review, Yale Journal on Regulation, Fordham Law Review, and Philosophy & Public Affairs.
Recent Posts
Recovering Strict Products Liability for the Age of AI
The rule of strict products liability in section 402A of the Restatement (Second) of Torts is widely assumed to have been formulated to address manufacturing defects. Relying on this assumption, the Restatement (Third) of Torts defines manufacturing defects as departures from the product’s design and therefore from the manufacturer’s manifest intent. Because virtually all product malfunctions stem from manufacturing defects, the Third Restatement extended this logic to define malfunctions as departures from manifest intent as well. This move effectively discards the section 402A consumer expectations test as a redundant relic from the early days of products liability.
Tort Law at the Frontier of Artificial Intelligence
This Article offers an in-depth conceptual, normative, and doctrinal examination of tort liability for foundation model development and release. It provides a qualified defense of the tort of negligence – the common law’s broadest and most flexible cause of action – as the principal doctrinal foundation of the tort system’s governance of this novel domain.
Insurability and Liability for AI-Caused Harms
The opacity of AI decision-making has led many tort scholars to conclude that ordinarily it will be infeasible to prove negligence or defect-based forms of products liability for AI-caused harms. According to mainstream tort theory, this evidentiary hurdle justifies strict enterprise liability for commercial AI distributors. Fully internalizing injury costs within these business enterprises adequately incentivizes them to adopt reasonably safe practices while relying on their liability insurance policies to efficiently and fairly compensate accident victims. Mainstream theory, however, decisively biases the analysis in favor of strict enterprise liability by not accounting for how the expansion of liability would substantially increase the cost of compensating injuries through insurance mechanisms.
Product Liability Law in the Age of AI
Although the title to this book has changed, this edition continues the developmental path staked out in the prior two editions of this casebook. When I began working on the liability issues involving the crashes of autonomous vehicles about a decade ago, I felt as if...
Andy Song on Civil Liability for Artificial Intelligence
An AI overview by Andy Song (Google Deepmind) from the 2025 Annual Meeting. Song delivers an engaging and accessible overview of the rapid advancements in AI technology, including the emergence of multimodal and “thinking” models capable of reasoning, planning, and...
Law-Following AI: Designing AI Agents to Obey Human Laws
Artificial intelligence (AI) companies are working to develop a new type of actor: “AI agents,” which we define as AI systems that can perform computer-based tasks as competently as human experts. Expert-level AI agents will likely create enormous economic value but also pose significant risks. Humans use computers to commit crimes, torts, and other violations of the law. As AI agents progress, therefore, they will be increasingly capable of performing actions that would be illegal if performed by humans. Such lawless AI agents could pose a severe risk to human life, liberty, and the rule of law.
The ALI Adviser is intended to inform readers about the legal topics and issues examined in many of ALI’s current projects; posts do not necessarily represent the position of the Institute taken in those projects. Posts on The ALI Adviser are written by ALI project participants, ALI members, and outside sources. Completed work is available to purchase online.
