President Trump Signs Scaled-Back Executive Order on AI Safety President Donald Trump has quietly settled on a compromise approach to artificial intelligence oversight, opting for a voluntary safety regime after weeks of internal fights over how much to restrain fast-moving AI companies.

From ambitious draft to private signing

In late May, Trump abruptly canceled a planned signing ceremony for a tougher executive order, warning it might become a “blocker” for US AI firms and jeopardize America’s edge over China. The earlier draft envisioned a 90‑day review window and broader federal authority to evaluate “frontier” models before release.

Industry allies and deregulation advocates inside the White House pushed back, arguing that long pre‑release reviews could dull “America’s edge on AI technology.” After this MAGA‑camp infighting, Trump delayed the order and solicited more input from tech leaders.

On June 2, he instead signed a narrowed order in a private setting, without the usual livestream or CEO‑lined photo op.

What the scaled‑back order does

The new directive creates a “voluntary framework” for AI companies to give the government access to powerful, cyber‑capable models up to 30 days before public release. It explicitly states that “nothing in this section shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement” for new models, including frontier systems.

Agencies are instructed to benchmark advanced cyber capabilities, stand up an AI cybersecurity “clearinghouse,” and treat AI‑assisted hacking and related crimes as enforcement priorities.

Competing views on risk and restraint

Supporters in the tech sector, including OpenAI cofounder Sam Altman, say the order “gets the balance right” between innovation and security, arguing the US should lead by building “the very best models” while “making sure they’re safe” and arming “trusted defenders” with cyber tools.

Critics, however, describe the move as a “watered‑down AI vetting order” that offers “performative reassurances” without changing how or when risky models are deployed. They question whether under‑resourced US security teams can meaningfully test cutting‑edge systems within a 30‑day window, especially after past budget cuts.

The result is a deliberately narrow order: a voluntary, time‑limited review process that leaves major decisions on hard AI rules for another day.

Continue reading https://foxvector.com/stories/019e8fc4-00bb-3f28-7027-146981e07a15