Key Regulatory Principles and Current Regulatory Approaches

The law and economics literature identifies the “judgement-proof problem” as a standard argument in law-making discussions operationalizing policies, doctrines and the rules. This chapter attempts to show that generative AI agent may cause harm to others but will, due to its judgement-proofness not be able to make victims whole for the harm incurred and might not have incentives for safety efforts created by standard tort law enforced through monetary sanctions. Moreover, the potential independent development and self-learning capacity of a generative AI agent might cause its de facto immunity from tort law’s deterrence capacity and consequential externalization of the precaution costs. Furthermore, the prospect generative AI agent might be employed by its users in ways designers or manufacturers did not expect (as shown in previous chapter this might be a very realistic scenario) challenges the prevailing assumption within tort law that courts only compensate for foreseeable injuries.