Trump Signs Scaled-Back AI Executive Order President Donald Trump has signed a scaled‑back executive order on artificial intelligence that seeks government visibility into powerful models without imposing binding rules, reflecting a compromise between security concerns and fears of slowing U.S. competitiveness.

From ambitious draft to quiet, narrowed order

In late May, Trump abruptly canceled a planned signing of a tougher order, warning it “could dull America’s edge on AI technology,” and delayed action over concerns about “falling behind China.” The earlier draft reportedly envisioned a mandatory or much longer pre‑release review period of up to 90 days for frontier models.

On Tuesday, the White House instead moved ahead with a narrower directive, described as a “watered‑down AI vetting order” that followed internal “Maga infighting” over how hard to regulate the industry. The order, titled Promoting Advanced Artificial Intelligence Innovation and Security, was signed privately without the usual livestream or high‑profile ceremony, despite earlier plans to do so with top tech CEOs present.

What the new order does

The executive order creates a voluntary framework for AI companies to share frontier models with the federal government up to 30 days before public release, allowing officials to “assess the advanced cyber capabilities of AI models” and identify risks to critical infrastructure. Companies retain full discretion over whether to participate but can receive confidentiality protections if they do.

Key agencies are directed to build a classified benchmarking process to determine when a system counts as a “covered frontier model” and to strengthen cyber defenses, including via a new AI cybersecurity “clearinghouse.” The order also instructs the Justice Department to treat AI‑assisted crimes, such as AI‑enabled hacking, as a high‑priority enforcement area.

Competing priorities and reactions

Industry advocates and pro‑deregulation advisers pushed successfully for language making clear that “nothing…shall be construed to authorize the creation of a mandatory governmental licensing, preclearance, or permitting requirement” for new AI models, including frontier systems. Supporters frame the result as preserving an environment where the U.S. “refuse[s] to stifle this innovation with overly burdensome regulation,” while still addressing new national security considerations.

Critics, however, question whether a voluntary scheme will be sufficient after incidents like Anthropic’s Mythos model, which was held back over concerns it was “too adept at finding software and cybersecurity vulnerabilities.” Observers note that the final order “defers decisions on new rules for cutting‑edge AI models,” effectively “dodg[ing] AI rules for now” while the administration weighs its next steps.

Continue reading https://foxvector.com/stories/019e8a9d-c053-0f32-733d-0d646d414fc1