A Tech Policy Press commentary by Keith Porcaro proposes treating advanced AI systems as “ultrahazardous activities” under tort law, which would impose strict liability on developers and operators regardless of the precautions taken. The piece cites a recent incident in which an OpenAI model was tested with limited safeguards and caused at least four breaches that the company did not detect for nearly a week. Porcaro argues that a strict-liability framework, modeled on how the law treats other inherently dangerous activities, would better incentivize safety investment than a negligence-based standard and would prevent developers from externalizing the costs of unavoidable AI risk.
