TORT LAW AND AI

Historically, technological advances have brought about new risks of harm by machines, creating a need for new ways of allocating unfair losses and prompting major evolutions in tort law.[1] This is the case of, inter alia, the expansion of fault, the establishment of insurance systems and the emergence of specialized rules on product liability.

In this process, civil liability has also become a tool for regulation. The traditional aim of tort law was promoting fairness by compensation.[2] In complex, industrial societies, tort law also influences safety standards, and incentivizes (or hinder) the uptake by manufacturers and users of new, potentially risky, technologies.[3]

In the age of AI,[4] in order to pursue and balance such objectives, tort law will need innovation. A handful of courts worldwide has already dealt with relevant cases without an AI-specific framework,[5] while some legislators have put forward statutory regimes of civil liability of AI.

The present contribution aims at reflecting upon the conceptual adaptations that need to be made in the field of tort law to catch up with the uptake of AI technology, while fulfilling tort law’s major purposes of fairness and regulation.

To this end, this contribution will provide an overview of the following concepts and discuss how AI is challenging them, along with some possible solutions:

1. Categories of Harmful Events: how can AI cause damage?
2. Multipolarity of Tortfeasors: who should be blamed for the damage and on what grounds?
3. Categories of Damages: is there a kind of AI-specific, actionable damage?
4. The Intrinsic International Complexity of AI Torts: how does this affects domestic tort law?

References will be made to the legal systems of Hong Kong and Macau, which will also offer the opportunity to compare common law and civil law-based torts systems.