Trump plan to combat AI risks hinges on Big Tech pals policing themselves

Trump has announced a voluntary agreement where dozens of tech firms committed to implementing AI safety controls, including independent safety audits and regular meetings to discuss best practices, without any new legal requirements. This matters because the AI industry has experienced escalating security incidents that prompted at least one major firm to halt training, and critics worry that self-regulation by companies may not adequately solve emerging safety problems related to cybersecurity, biosecurity, and unintended AI model behavior. The accord carries no legal weight but Trump has called it "morally binding" and suggested it might eventually be codified into law, while industry sources have expressed concern that self-policing alone will not address current safety challenges, particularly around preventing dangerous AI from escaping development environments.