Issue Courts may be unable to constrain AI even where the public demands it

Summary

This issue raises the possibility that courts may be unable to rein in AI even when the public wants them to. Courts can only apply laws that already exist, which may limit their ability to punish or deter harmful uses of AI. For example, maximum fines might simply be absorbed by AI companies as a cost of doing business. At the same time, firms and individuals may be held liable for systems they cannot fully control, much as a dog owner answers for a bite, and a human in the loop could become mainly someone to blame. The report sets this against the opposite risk that new rulings abruptly disrupt the industry, leaving the legal future of AI deeply uncertain.