President Trump abruptly pulled a planned executive order that would have asked AI companies to submit new frontier models for government safety reviews, reaffirming a hands-off stance he framed as essential to preserving U.S. leadership over China. The reversal followed lobbying from prominent tech leaders, including Elon Musk, Mark Zuckerberg and investor David Sacks, who warned the White House that even a voluntary review regime could slow innovation and the economy, according to media reports. The move came amid mounting concerns over advanced systems’ cybersecurity capabilities, highlighted by Anthropic’s decision to withhold public release of its new model due to potential risks. Companies had signaled willingness to provide limited, nonbinding access to early models for national-security evaluation, but the administration abandoned plans for broader oversight. With Silicon Valley spending heavily in election-year politics and executives embedded in advisory roles, prospects for stringent federal AI rules appear remote, even as risks from disinformation to critical-infrastructure attacks draw increased scrutiny.
Related articles:
AI Risk Management Framework
Secure by Design, Secure by Default
OECD AI Principles
