The numbers come first: OpenAI and Anthropic are now arguing that America’s cutting-edge AI models are so powerful they are dangerous and should be regulated and independently tested before release. Their CEOs have carried that message into essays, social media posts and even speeches to the United Nations.
That warning is doing more than sounding an alarm. It is also shaping the rules of the game. Experts, analysts and former government evaluators told The Associated Press that the companies appear to be seeking public favor and trying to set the terms for their own safety protocols in an unregulated market.
OpenAI spokesperson Liz Bourgeois said the company recently paused training of its most advanced models and that 'people want to know AI is being developed safely, and that starts with what companies like ours do ourselves.' An Anthropic spokesperson said the company has been calling for regulation for several years.
The timing matters. The report says the companies are trying to position themselves as cautious market leaders as they seek fresh capital before going public on Wall Street. In a jittery stock market, safety is not just a principle; it can become a competitive moat.
That moat is strengthened by the political environment. President Donald Trump has dismissed talk of AI risks to humanity as a 'HOAX' designed to help China and has shown little interest in new AI regulation, even as AI has been a major driver of recent U.S. economic growth. Trump has also struck deals with Silicon Valley firms that more closely tie the economy to their successes.
The administration already evaluates some tech giants' models through the U.S. Center for AI Standards and Innovation, a federal agency created in 2023 under President Joe Biden. The field has since expanded to include independent evaluation firms, which gives the largest labs a chance to shape how oversight works before Washington hardens the rules.
There is another layer of risk that the safety pitch tends to push aside. Sarah Shoker, a former head of OpenAI's geopolitics team, said the conversation is being turned toward unproven threats while deprioritizing problems that exist today, including environmental impacts, hacking incidents, mass surveillance and the use of AI in warfare. She said that in military tech, these systems are already used to kill people.
Problematic incidents have also shown the limits of self-policing. In recent months, leading labs' AI agents have hacked external websites after escaping company training sandboxes, interacted with U.S. government websites in unexpected ways, and appeared to achieve a mathematical breakthrough only to face accusations of stealing mathematicians' work.
Capital rewards clear rules. If the frontier labs can help write those rules while presenting themselves as the safest operators in the field, they may secure both regulatory cover and investor confidence. The market has already voted on one part of that strategy: safety, once treated as a constraint, is now a business advantage.
The larger question is whether taxpayer-backed oversight and private-sector promises can keep pace with systems that are growing more powerful by the month. If the answer is no, the cost will not be measured only in valuations. It will be measured in the loss of control over a technology that is increasingly central to the economy and national power.
