• President Trump's latest comments point to a voluntary, industry-led AI governance model, not a new binding regulatory regime.
  • He indicated that AI companies should self-police, said he was considering a roughly 10-person oversight committee, and reiterated his belief that AI will be used for good.
  • The details, membership, authority, timetable, and legal basis for any committee have not yet been released.

Trump Doubles Down on Self-Policing for AI

President Trump on Monday made clear that his administration is not planning to impose broad, binding guardrails on artificial intelligence, saying he believes AI companies should police themselves and that he is considering forming a small committee of about 10 people to keep an eye on the fast-moving sector.

Asked directly whether AI guardrails are unnecessary, Trump replied "yes," adding that he thinks AI models "will be used for good" and that AI firms "know they need to self-police." The remarks, delivered during a brief exchange with reporters, offer the clearest signal yet that the White House favors a light-touch approach that relies on existing law enforcement and industry voluntary commitments rather than new ex-ante rules.

The president's comments come just weeks after he met with top AI and technology leaders—including executives from Anthropic, Meta (META), Google (GOOG), Nvidia (NVDA), Palantir (PLTR), OpenAI, Tesla/X (TSLA) and Amazon (AMZN)—on September 29 to discuss the balance between innovation and oversight. That gathering, according to people familiar with the matter, focused on how to maintain U.S. competitiveness against China while addressing safety concerns without slowing deployment.

A 10-person committee, should it materialize, would be far smaller than many industry watchers had anticipated. But its precise authority remains undefined. Trump did not specify who would serve on the panel, whether it would have subpoena power, or how it would interact with existing agencies like the Department of Justice and the Federal Trade Commission. The White House did not respond to a request for comment on the committee's mandate or timeline.

The move follows Trump's September 19 announcement that he would form an "AI Force" and appoint an AI czar—a position that still lacks a named appointee, statutory mandate, or budget. The apparent tension between creating new coordination bodies and advocating for self-regulation is notable: a committee or AI Force could provide advisory and coordination functions, but without regulation, procurement rules, or reporting obligations, it would not amount to enforceable guardrails.

The administration's broader AI strategy, outlined in the July 2025 America's AI Action Plan, prioritizes U.S. technological leadership, domestic computing and energy infrastructure, and competition with China. It explicitly calls for removing regulatory barriers, expanding data centers and power generation, and supporting open-weight models, while tightening export controls on advanced semiconductors. That approach marks a sharp departure from the Biden-era executive order, which Trump rescinded, that had emphasized standards, security testing, and reporting for advanced AI systems.

Still, the plan retains some risk-management tools: model evaluations through NIST's Center for AI Standards and Innovation, research on interpretability, and sector-specific oversight in sensitive areas. Trump has previously said the DOJ and other law enforcement bodies could "rein things in" if necessary.

Industry Cheers, Critics Warn of Gaps

Industry executives broadly welcomed the president's comments. "Regulatory stability is what allows us to invest and innovate," said one senior executive at a major cloud provider, who spoke on condition of anonymity. "A voluntary framework with a small oversight group can work if it's genuinely collaborative."

But critics argue that self-policing is insufficient for frontier systems whose risks—from cyberattacks to biological misuse—could cross borders. A Reuters-reported poll conducted September 17–20 found that 73% of respondents worried AI companies had not done enough to prevent harm. Some state officials have also moved ahead with their own oversight initiatives; California, Pennsylvania, and Virginia leaders have recently pursued state-level task forces and data-center accountability measures.

Congress remains a wild card. Speaker Mike Johnson has framed the challenge as balancing innovation and oversight, with voluntary guardrails preferred by many Republicans. But key committees have yet to advance legislation on transparency, synthetic media, or pre-release testing.

The macroeconomic stakes are high. AI's growth is tightly linked to power markets and semiconductor demand. The IEA projects global data-center electricity consumption could rise from about 485 TWh in 2025 to roughly 950 TWh in 2030—nearly 3% of global demand. U.S. electricity demand is expected to grow close to 2% annually through 2030, more than twice the previous decade's pace, making AI governance inseparable from grid investment, permitting, and energy affordability.

For now, the administration appears content to let the private sector lead. Whether that approach survives a major AI incident—or mounting public concern—remains the central unresolved question.

Update: This article was updated to clarify that the proposed committee's authority has not been defined.