Former OpenAI employee says AI should be regulated like nuclear power plants

David Robinson, a former safety lead at OpenAI, has resigned and warned that the company lacks sufficient care regarding AI risks. In an essay for The Atlantic, Robinson argued that frontier AI model releases require layers of redundancy and meticulous planning. He suggested that AI labs should be regulated similarly to nuclear power plants, criticizing the trial and error approach that has historically driven the growth of the ChatGPT maker to ensure better safety standards.