Washington – In a move that underscores the Trump administration’s pragmatic approach to emerging technology, U.S. AI developers may now provide early‑stage models to Britain’s AI Security Institute for safety testing without violating federal guidance. The change, reported by the Telegraph, reverses an earlier request that companies such as OpenAI and Anthropic refrain from sharing unreleased systems with the British agency.

Why the shift matters

The United Kingdom established its AI Security Institute after the world’s first AI safety summit in 2023. The institute’s purpose is to evaluate the capabilities of cutting‑edge models and assess the risk of misuse before they reach the public. By allowing U.S. firms to cooperate with the institute, the administration is supporting international collaboration on AI safety while preserving American innovation.

Administration’s rationale

Spokespersons for President Trump emphasized that the updated stance reflects a “softer, more cooperative approach” to global AI governance. The White House argues that sharing information with trusted foreign partners helps identify vulnerabilities early, reducing the chance of dangerous misuse. This aligns with the administration’s broader AI task force, which recently required developers to report security incidents involving their models.

Industry reaction

OpenAI and Anthropic have not yet commented, but the policy change is expected to give them greater latitude to engage with allies on safety standards. Industry observers note that coordinated testing can accelerate the development of safeguards without stifling progress.

Balancing safety and innovation

President Trump has previously questioned the need for binding safety mandates on AI developers. The current adjustment suggests a nuanced shift: while the administration remains skeptical of heavy‑handed regulation, it recognizes the value of voluntary, cross‑border cooperation to protect the public.

What’s next?

The AI task force will continue to monitor compliance and security reporting, ensuring that any incidents are promptly addressed. By fostering a collaborative environment with Britain’s AI Security Institute, the United States aims to stay ahead of potential threats while maintaining its leadership in artificial‑intelligence research.

Spokespersons for the White House, OpenAI and Anthropic were not immediately available for comment at the time of publication.

Original reporting: Appleton, WI News Feed (HLL/CB) — read the source article.