Trump Cancels Signing of AI Executive Order President Donald Trump’s last‑minute decision to shelve an AI safety executive order has exposed a sharp split between efforts to rein in powerful models and a White House wary of slowing U.S. tech dominance.

The process began as officials, alarmed by cyber‑capable systems like Anthropic’s Mythos and OpenAI’s GPT‑5.5 Cyber, drafted an order to tighten cybersecurity and give the government early access to “frontier” AI models. The plan, reported in mid‑May, would bolster protections for national security agencies and critical infrastructure while creating a voluntary regime for companies to share advanced models up to 90 days before release.

Within the administration, concern grew over the political fallout if an AI‑enabled cyberattack hit the U.S., prompting work on an AI oversight order that could grant government first access to new systems without outright blocking them.

By May 21, the White House had scheduled a high‑profile signing. But behind the scenes, Trump adviser David Sacks and several industry voices argued the order went too far, with one source saying Trump “just hates regulation” and calling the effort “something doomers wanted.” Meta’s Mark Zuckerberg and xAI’s Elon Musk also spoke with Trump in the hours before the event, as industry critics warned pre‑release testing could delay launches and force costly redesigns.

Trump ultimately postponed the order, saying he “didn’t like certain aspects of it” and feared it “could have been a blocker” for jobs and the “tremendous good” AI is creating. He repeatedly cited competition with Beijing: “We’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that leading.”

The photo‑op signing was then abruptly canceled outright after several top AI CEOs declined to attend on just 24 hours’ notice, leaving others “midair on their way to the Oval Office” when they learned the trip was for nothing. With the order on ice, both the administration and the AI industry are now “scrambling to figure out what’s next” for frontier‑model oversight and cyber risk, and it is unclear whether any AI safety framework will emerge soon.

Continue reading https://foxvector.com