CHAPEL HILL, N.C. — Officials from the Trump administration attended this week’s G20 Innovation Ministerial aiming to project American leadership in artificial intelligence, but tensions between the Commerce Department and the White House Office of Science and Technology Policy (OSTP) surfaced behind the scenes.
Why it matters
The United States is urging other nations to follow its lead on AI policy, favoring lighter-touch approaches versus stricter frameworks such as the European Union’s. At the same time, internal disagreements complicate Washington’s ability to speak with one voice about oversight of the most advanced AI models.
What happened at the summit
The G20 Innovation Ministerial took place on Tuesday and Wednesday and gathered senior industry figures and government officials from the world’s largest economies. U.S. participants framed the country as a model for encouraging private investment through “flexible policy frameworks,” a line emphasized by OSTP Director Michael Kratsios.
The institutional struggle
Commerce and OSTP are competing for influence over the administration’s AI agenda. Key regulatory choices remain unresolved while the U.S. seeks to present clear guidance internationally.
A White House spokesperson dismissed reports of disarray as “fake news,” saying the president’s team is “working closely together to showcase American leadership” and that “Commerce and OSTP are jointly leading a productive and successful G20 Innovation Summit.”
Visible differences in roles and programming
The first day of the summit highlighted differences in how OSTP and Commerce view their roles and each other. Commerce controls several of the administration’s main regulatory levers on AI, including the Bureau of Industry and Security and the Center for AI Standards and Innovation. One source familiar with the administration described OSTP as a shop of “thinkers” that lacks the regulatory authorities held by Commerce.
The programming reflected that split: Kratsios led Tuesday’s agenda and took part virtually with Elon Musk, Demis Hassabis (chair of Google DeepMind) and Mark Zuckerberg (Meta). Commerce Secretary Howard Lutnick was scheduled to headline Wednesday, hosting in-person fireside chats with Jensen Huang (Nvidia), Sam Altman (OpenAI), Tom Brown (Anthropic) and Alex Karp (Palantir).
Lutnick told Axios he prioritized in-person participation: “I tried to bring deeply interesting people, but only in person… If you're not gonna get on the plane and come and join where the top 20 countries of the world are, that's OK.” Kratsios told reporters “We’re one team.”
A source said Musk might have attended had pre-summit coordination been better; another source said Musk had been confirmed as a virtual participant for three weeks and his team consistently chose that format.
Data centers became a flashpoint
Coordination gaps spilled into messaging on data centers, a politically sensitive subject for the administration. Kratsios raised the issue in his conversation with Musk even though officials had planned to avoid it during the summit after President Donald Trump’s Truth Social post reading “let Data Reign,” a person familiar with the administration’s thinking said. That source said officials wanted to avoid the topic because it has become a midterm flashpoint and Chinese representatives were in the audience.
Kratsios later echoed Trump’s position to reporters, calling decisions about data centers ultimately a “community decision” that can deliver local benefits if providers supply their own power.
The broader picture on AI oversight
The U.S. is pressing other countries to delay imposing new AI regulations while internally it continues to debate how to oversee frontier models. The White House has finalized—but not publicly released—a framework for reviewing certain frontier models before deployment.
The administration remains divided on major elements of its proposed approach, including a FINRA-like regulatory body for AI and a replacement for the Biden-era AI diffusion rule that governed exports of advanced chips and models.
Statements ahead of the summit
Ahead of the meeting, a White House official told Reuters the U.S. would “press member countries not to set up new regulatory organizations to oversee AI development.” David Sacks, an external adviser to Donald Trump on AI, said in a virtual address to G20 ministers that he opposed creating a new regulatory body, arguing there is already “a thicket of laws that apply to AI.”
Bottom line
Debates over who sets and controls AI policy in Washington are unfolding in parallel with U.S. efforts to shape global AI governance. Internal rivalries and mixed messaging at the G20 highlight that domestic power struggles may influence how persuasive America’s call for lighter regulation will be.
