Trump administration moves to test AI models and consider equity stakes in US AI firms

An executive order signed by President Donald Trump directs major AI developers, including OpenAI, Google and Anthropic, to voluntarily submit their most advanced models for a 30‑day government cybersecurity review before public release. The testing program, coordinated by agencies such as Homeland Security, Defense, Commerce and Treasury, aims to identify vulnerabilities that could affect critical sectors like banking, healthcare and emergency services.

Separately, senior officials in the administration have held early‑stage talks with leading AI companies about the possibility of the federal government acquiring voluntary equity stakes in those firms. OpenAI CEO Sam Altman has championed the idea, which would channel dividend payments from any government‑held shares to American households. The discussions are preliminary, face legal and conflict‑of‑interest challenges, and come as companies like OpenAI and Anthropic prepare for large IPOs.