brief voluntary – President Donald Trump has privately signed a scaled-back executive order that gives AI and tech companies up to 30 days to voluntarily allow the federal government to review AI models with advanced cyber capabilities before public release. The order comes aft
President Donald Trump signed a scaled-back executive order that creates a short, voluntary window for the federal government to review AI models before they go public.
Under the new terms, frontier AI model makers and tech companies can choose to provide the federal government access to AI models with advanced cyber capabilities up to 30 days before a public release. None of the reviews are mandatory, according to the order’s emphasis.
The timing is the part that has technology companies paying attention. A previous version of the executive order—published by Politico—had called for a review period of up to 90 days. This time, the window has been tightened, and the reviews remain optional.
The move lands after repeated federal conversations with leading AI companies about cybersecurity, including Anthropic. In April, Anthropic said it was limiting the release of Claude Mythos because it was concerned the model was “too adept at finding software and cybersecurity vulnerabilities.”
Anthropic, which has confidentially filed for an initial public offering, said last week it would soon release Mythos-level models to its customers.
The order also lands in the middle of Trump’s broader argument about US competition in AI. On May 21. Trump told reporters he had delayed signing the order over concerns it could endanger US companies’ lead over China in the AI race. “We’re leading China. We’re leading everybody,” he said. “And I don’t want to do anything that’s going to get in the way of that lead.”.
The revised executive order makes the trade-off concrete: the government gets a chance to examine advanced cyber-capable models. but companies are not required to hand them over. and the review window is shorter than the earlier draft. For startups and established AI labs alike. the difference between 30 days and 90 days can change release schedules—and the balance of control between innovation and oversight.
Trump executive order AI models cybersecurity federal review Anthropic Claude Mythos initial public offering China AI race
So it’s voluntary but still “review”? sounds like a loophole parade.
I don’t get how 30 days is enough to check anything cybersecurity related. If they’re “leading China” then why even delay and then sign this. Seems like politics not tech.
Wait, Anthropic said Claude Mythos was too good at finding vulnerabilities, but then they’re gonna release it to customers anyway? That’s basically admitting the AI is a hacking tool and just putting a calendar on it.
This is just like when they “review” medicines for 30 days and then say it’s safe lol. Voluntary means they’ll only show the government the nice parts of the model. Also why is China always the excuse, can’t we just do regulation without all the race talk.