Closing the AI Accountability Gap: Strict Liability and Punitive Damages for Advanced Artificial Intelligence

Abstract

Advanced artificial intelligence can generate physical-harm risks whose extreme downside may be practically non-compensable—losses can exceed insurable limits, bankrupt defendants, or arise in scenarios where the legal system cannot meaningfully function. Existing U.S. tort doctrine is likely to underinternalize these low-probability, high-consequence risks. Fault-based standards ask juries to assess “reasonable” precautions and reasonable-alternative-design questions under deep technical uncertainty; scope-of-liability limits may push emergent behavior or foreseeable misuse outside liability; and punitive damages generally require human malice or reckless disregard. 

This Article proposes a tort framework aimed at closing that accountability gap. First, for a defined subset of frontier systems, it argues that training or deploying them should be treated as an abnormally dangerous activity, triggering strict liability for physical harms within the risk that makes the activity dangerous. Second, it develops a deterrence-based approach to punitive damages for compensable “warning-shot” injuries when there is precaution overlap: when the safety measures that would have reduced the plaintiff’s injury risk would also tend to reduce an associated uninsurable downside, courts should be permitted to award punitive damages calibrated to internalize the expected value of that downside. The Article sketches administrable evidentiary approaches to overlap (including capability, misuse, and alignment evaluations), addresses constitutional due process constraints, and evaluates legislative complements such as capability-scaled liability insurance.

Keywords: AI, artificial intelligence; torts; existential risk, catastrophic risk, law and economics, externality, strict liability, punitive damages, liability insurance, AI governance