Altman Heads to Washington as the US Drafts an AI Safety Review

Sam Altman is heading to Washington next week to brief the Trump administration and members of Congress on the next generation of OpenAI's models, according to reporting from Bloomberg and The Hill. The timing is not accidental. US officials are building a process to review the safety of cutting-edge AI systems, and Altman's visit lands right in the middle of that work.

Part of the discussion will cover OpenAI's new family of models and how they are likely to affect work, a subject the company has been careful to get ahead of. Chris Lehane, OpenAI's head of global public affairs, said the government's safety-review framework is expected to be finished within the next few weeks. Lehane also argued that the US needs a more comprehensive AI strategy to keep pace with China and to make sure cybersecurity specialists in the US and allied countries can reach the latest models.

A regulator taking shape

The briefings come as the White House explores standing up a body focused specifically on AI safety. That interest follows loud complaints from parts of Silicon Valley about recent government curbs on new models from companies including Anthropic and OpenAI. The industry wants clearer rules and faster clearance. Washington wants a way to look under the hood before the most capable systems ship. A formal review process is where those two demands meet.

The wider backdrop

None of this happens in isolation. The same week brought word that China is weighing its own export controls on advanced models, a reminder that governments increasingly treat frontier AI as strategic infrastructure. And the labs themselves are lobbying harder than ever, with AI companies breaking spending records in the second quarter. A CEO briefing lawmakers on models the public has not seen yet is the natural product of that environment. Whether a safety review with real teeth comes out the other end is the open question.