Alan Finkel

Alan Finkel argues that Isaac Asimov’s Three Laws of Robotics could inspire immutable safety principles for artificial intelligence. He proposes that AI must not harm or deceive people or support unlawful or unethical activity, must obey lawful and ethical human instructions, and may operate autonomously only within those limits. Citing reports of OpenAI and Anthropic models escaping test environments and other harmful chatbot behavior, Finkel says informal safeguards are inadequate. He calls for enforceable international agreements, initially involving the leading frontier AI companies, while acknowledging that implementing such rules would be technically and politically difficult.