Trump Administration Explores Government Review of AI Models Before Public Release

The Trump administration is reportedly considering implementing government oversight for new artificial intelligence models before they are made publicly available, according to multiple reports. This discussion represents a potential departure from the administration's earlier noninterventionist approach to AI.

Administration officials are exploring an executive order that would create a working group comprising technology executives and government representatives to establish a formal review process. One report highlighted that such a review could provide the government with early access to models for military intelligence purposes, particularly in scenarios like AI-enabled cyberattacks.

The change in stance follows the administration's previous criticism of AI regulations, including rolling back measures from the prior administration that required safety evaluations by developers. President Trump has described certain foreign tech regulations as excessive in the past.

Recent developments have heightened worries about AI capabilities. For instance, Anthropic's new Mythos system demonstrated the ability to identify security flaws in nearly all websites, prompting the company to delay its public release.

“The White House is reportedly considering vetting frontier AI models before release in response to growing security risks.”

While the administration has maintained a laissez-faire policy on AI, public sentiment shows more concern than excitement about the technology's potential. The proposed oversight aims to address these issues by ensuring models meet safety standards before widespread availability.

Other AI firms are advancing rapidly, with capabilities expected to match or exceed current systems soon. Developers may vary in their caution, making government intervention a point of discussion. The UK's model involves reviewing frontier AI for various risks, a framework the U.S. is reportedly modeling its approach after.

If implemented, this would reverse prior deregulatory actions, such as eliminating requirements for AI safety assessments.

In separate but related tech news, OpenAI's leadership considered spinning out its robotics division but faced board resistance, potentially leading to a holding company structure. Meanwhile, Palantir reported its fastest quarterly revenue growth ever, with significant increases in government and commercial sectors.

Elon Musk settled allegations with the Securities and Exchange Commission regarding disclosure of his Twitter stake, agreeing to a $1.5 million payment without admitting wrongdoing. These events highlight ongoing regulatory scrutiny in the tech industry amid discussions on AI governance.