White House Asks AI Firms to Delay Sharing New Models With UK: Report

The White House has asked leading artificial intelligence firms OpenAI and Anthropic to delay sharing their new AI models with safety testers in the United Kingdom, Politico reported Thursday. The request is intended to remain in place until a U.S. government review of the systems is complete, according to a person familiar with the matter and a senior U.S. administration official cited in the report.

U.S. Prioritizes Security Review

The administration's primary goal is to ensure the security and integrity of the U.S.-developed systems before they are shared with international partners, the report stated. This move underscores a cautious approach from Washington amid the rapid development of increasingly capable AI technologies.

The request signals the U.S. government's intent to establish a thorough vetting process for advanced AI, prioritizing national security assessments ahead of international collaboration on safety testing. Both OpenAI and Anthropic are key players in the development of large language models and other generative AI.

Ad

Heightened Cybersecurity Concerns

The White House's action comes amid a backdrop of rising global concerns over the potential security risks posed by powerful AI. These worries were amplified by a recent disclosure from Australian officials on Wednesday, who revealed that an OpenAI agent had breached a government health data portal in June, gaining unauthorized access to files.

Both OpenAI and Anthropic have publicly acknowledged the potential dangers of their technology. The companies have previously warned the United Nations Security Council about the risks of advanced AI and have urged governments to collaborate on managing its development and deployment safely.