President Donald Trump reportedly declined to create an industry-funded AI standards body after Meta CEO Mark Zuckerberg, Nvidia CEO Jensen Huang and SpaceX CEO Elon Musk separately argued against the proposal, The Wall Street Journal reports. The intervention is the latest turn in an unresolved fight among Trump advisers and technology executives over how closely the federal government should scrutinize advanced AI systems.
Google DeepMind co-founder Demis Hassabis had pitched a regulator modeled on the Financial Industry Regulatory Authority, the private organization that oversees U.S. brokerages. The Journal says Zuckerberg, Huang and Musk raised concerns about who would control the new body and whether it would concentrate power among OpenAI, Anthropic and Google. Trump ultimately did not agree to establish it. The Associated Press has separately described Hassabis's proposal as an industry-funded group that would develop testing standards for national-security risks.
Two camps around Trump
Inside the administration, Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent and National Cyber Director Sean Cairncross have pushed for greater scrutiny, according to the Journal. Wiles, Bessent and other officials have reportedly met almost daily in recent months to discuss AI risks. Bessent has focused in part on the danger of AI-enabled attacks on the banking system, while officials have also examined threats to water, power and other critical infrastructure.
On the other side, White House technology adviser David Sacks, Zuckerberg and Huang have generally pressed Trump to keep federal involvement limited. The Journal reports that company executives have sometimes bypassed the broader policy process and contacted the president directly, adding to frustration among officials trying to build a consistent approach.
A narrower order survived the earlier fight
The same divide shaped an executive order earlier this year. Sacks persuaded Trump in May to drop a more demanding version after a last-minute call, according to the Journal. Wiles and Bessent later persuaded him to sign a narrower measure on June 2.
The final executive order directed the Commerce Department to develop classified benchmarks for the cyber capabilities of frontier AI models. It also created a cybersecurity clearinghouse designed to work voluntarily with AI developers and critical-infrastructure operators. Axios reported in August that the implementation framework covers certain closed frontier models while excluding open models.
