OpenAI Wants AI Rules, White House Calls It a Hoax

Two Messages in One Week

OpenAI told Congress it needs mandatory AI safety rules, and five days later the president told Nvidia's chief executive the entire risk is a hoax. Both statements happened inside a single week in September 2026, and both came from people whose decisions shape what happens to the AI systems a European or British business now runs on. For a company choosing between GPT, Claude and Gemini as a platform, the regulatory ground under all three vendors moved in opposite directions at the same time.

The week opened with OpenAI publishing a policy essay asking Washington for binding oversight it had resisted until now. It closed with Donald Trump calling that entire debate fake, live onstage, to the head of the company that sells the chips every one of these labs depends on. Between the two, Anthropic's chief executive published a plan to slow the whole industry down, and three rival lab heads signed their names to it.

What OpenAI Actually Asked For

Chris Lehane, OpenAI's chief global affairs officer, published a policy post on September 9 asking Congress for mandatory, capability-based national regulation, not the voluntary commitments OpenAI has relied on until now. The post is blunt about the reason: "The United States needs mandatory, capability-based national regulation that can evolve as the technology does."

The framework OpenAI wants includes common testing and independent-assessment requirements, stronger cybersecurity protections, clear incident-reporting rules and shared measures for tracking progress toward recursive self-improvement, the point at which an AI system starts driving the design of its own successor. Lehane is explicit that this should apply narrowly, to a handful of well-resourced labs building the most capable systems, not to startups or open-source developers.

Until Congress moves, OpenAI is also endorsing four California bills headed to Governor Newsom: SB 813 on independent safety assessors, AB 1405 on auditor accountability, SB 1119 on protections for teenagers using companion chatbots, and AB 1864 on screening for AI-enabled biological threats. OpenAI calls this approach reverse federalism, where state laws set a de facto baseline Congress can later codify. Some of these bills OpenAI opposed before; the post says the company changed position after what it calls a recent jump in model capabilities.

Amodei's Three-Step Plan

Anthropic's chief executive Dario Amodei published a roughly 3,800-word essay on September 12 titled We Must Pace the Frontier, and three rival lab heads signed on within two days. The essay proposes coordinated pacing at three levels: third-party evaluators with employee-level access inside frontier labs, common safety standards among AI companies in democratic countries, and an attempt at coordination with authoritarian governments on the pace of development itself.

Amodei's stated concern is recursive self-improvement outrunning human oversight, alongside cyberattacks and bioterrorism risk from more capable models. He is careful to frame the plan as protecting, not sacrificing, US advantage: pacing "would give frontier AI developers the time to do this vital work without sacrificing commercial advantage or the United States' lead in AI." On China, he is equally blunt about the danger of unilateral restraint: "If we greatly restrain our AI capabilities in the belief that China will do the same, and then China defects," any agreement needs either verifiable enforcement or a scope narrow enough that defection would not hand China a military edge.

Sam Altman, Google DeepMind's Demis Hassabis and xAI's Elon Musk all cosigned the essay within days, an unusual public alignment among CEOs who normally compete on how fast, not how cautiously, their labs move.

The White House Calls It a Hoax

President Trump interrupted the All-In Summit in Los Angeles on September 14 to call Nvidia chief executive Jensen Huang onstage and declare the entire AI safety debate fake. "The robots will not be taking over. The AI will not be taking over the rest of the world. The whole thing is a hoax," Trump told Huang, adding that opposition to data center construction helps geopolitical rivals such as China and that AI infrastructure will be "the oil of the next 20, 25 years." Huang did not push back.

The call landed two days after Amodei's essay and the same week OpenAI asked Congress for binding rules, staking out a clear split between the AI industry's own safety leadership and the administration whose Congress would have to pass any federal framework. The FRONTIER Act, a bipartisan House bill that would require large developers to undergo independent audits and report critical safety incidents, is moving through committee at the same time, meaning the legislative fight and the political fight over whether AI risk is even real are now happening in parallel.

What This Changes for a Vendor Risk Review

None of this changes what the EU AI Act already requires of a business deploying GPT, Claude or Gemini inside the EU, but it changes how much that business can assume about the US regulatory backdrop those vendors operate under. A month ago, a reasonable planning assumption was that federal AI safety legislation was gathering bipartisan momentum behind bills like the FRONTIER Act. That assumption no longer holds when the sitting president has just called the underlying risk a hoax in front of the CEO whose chips the entire industry needs.

The practical shift is where the near-term rules will actually come from. OpenAI's own reverse federalism bet means California, not Washington, is the jurisdiction to watch for the next binding requirements on frontier labs, through SB 813 and the three bills alongside it now sitting on Governor Newsom's desk. A business that wants to track vendor obligations in real time is better served by following those state bills and each lab's own published safety framework directly than by waiting for a federal outcome that no longer looks close.

Servola Journal

We do this for everyone trying to keep up with what technology is doing to our lives. The people who build it, and the people it happens to. The Servola Journal exists so that what we learn belongs to all of them.

Nobody pays us for this. No ads, no paywall, free to everyone. We just believe that understanding what's happening to all of us shouldn't depend on who can afford to pay for it.

If it gave you something today, tell us to keep going. Follow us, leave a like, or write a positive comment. We read every one, and they are what keeps us going.

Read next: Nvidia Fell 3 Percent. Trump Called Its CEO Onstage.  |  Three AI Labs Are Quietly Writing Their Own Rulebook