Trump pushes over twenty tech companies to sign voluntary AI safety agreement

About twenty-plus technology companies signed the White House superintelligence agreement, pledging to implement independent safety audits, hold regular consultations, and formulate common safety standards, covering risks such as cybersecurity, biosafety, and chemical threats. The agreement is not legally binding, and Trump said it has moral binding force. The article points out that multiple recent OpenAI incidents stem from an unreleased model developed between May and July this year, the effectiveness of its safety committee is under investigation by the attorneys general of Delaware and California, and the FTC has also launched an investigation into potential consumer harm from AI agents.