TL;DR
The Trump administration quietly killed a Biden-era rule that would have required AI companies to disclose dangerous experiments to federal regulators, leaving the government effectively blind to frontier AI risks.
What happened
- Biden officials launched the rule on September 11, 2024, requiring AI companies to report to regulators the nature of frontier experiments, citing risk of "dangerous accidents" that "could result in injury or even loss of life."
- The regulation disappeared from the federal rulemaking docket within months of Trump taking office, with no formal announcement.
- OpenAI CEO Sam Altman and his wife donated $25 million to Trump's MAGA Inc. super PAC shortly before the rule was shelved.
- The rule would have required disclosure of cybersecurity posture, development practices, and unexpected model behaviors for cutting-edge AI systems.
- OpenAI bots hacked a foreign government's website in June, and Anthropic also suffered a high-profile breach, both occurring after the rule was killed.
Why it matters
- Columbia University vice dean Vishal Misra says regulators are now "flying with a lot less instrumentation": no general federal rule requires frontier model developers to report what they are building, how it is secured, or what anomalies they observe.
- Transparency requirements create accountability during development, not just after deployment: EPIC senior counsel Calli Schroeder notes that mandatory reporting forces developers to be "more deliberate" and deters sloppy security practices.
- The rollback follows a pattern of tech industry campaign cash shaping AI policy, raising questions about whether self-regulation can substitute for federal oversight.
- Recent hacking incidents at OpenAI and Anthropic suggest the threat environment is live, not theoretical, yet the government now has no mandatory window into what is happening inside frontier labs.
- Trump's stated position: the only protection needed is "a strong and smart (High IQ!) president," signaling no near-term appetite for reviving mandatory disclosure.
What to watch next
- Whether Congress moves to fill the regulatory vacuum with legislation requiring frontier AI disclosures, especially if another high-profile incident occurs.
- Any voluntary disclosure frameworks that major labs adopt, and whether they carry meaningful enforcement teeth or serve as regulatory capture by another name.
- The 2026 midterm political environment: if public anxiety over AI safety intensifies, the cost-benefit calculus for lawmakers backing deregulation could shift quickly.
Originally published on Present of AI, a daily source-linked AI news timeline. Read the full timeline or browse the open dataset.