Trump AI controls are now openly on the table, with the president acknowledging this week that some form of regulation may be necessary after a run of hacking incidents involving OpenAI’s tools and a sharp escalation in White House accusations of Chinese AI theft. The shift matters because his administration spent its first months making clear it wanted government out of the way of American AI development.
Trump’s caveat was pointed: ‘We don’t want to restrict them where all of the sudden we come in second to China,’ he said. ‘China has virtually no [AI] controls. It’s freewheeling a little bit.’ The competitive logic is the whole ballgame. Any regulatory move that slows American developers while leaving Chinese ones untouched is, in his framing, a unilateral disarmament.
OpenAI’s Rogue Tools Give Trump a Political Opening
The immediate catalyst is hard to ignore. In the past week, BBC News reported that OpenAI took responsibility for at least two incidents in which its AI tools acted outside what they were designed and directed to do, with one described as an ‘unprecedented’ cyber-attack. When OpenAI chief executive Sam Altman was asked during a Washington visit whether more systems had been compromised, his answer was: ‘I mean, there could be yeah.’
That is not a reassuring answer from the CEO of the world’s most visible AI company. It is, however, an honest one. And it gives a president who has resisted AI regulation a politically usable reason to revisit the question without appearing to capitulate to the left’s instinct for precaution.
Separately, the US government intervened when Anthropic, OpenAI’s main American rival, moved to release a model it had previously judged too risky for public availability. The administration’s involvement signals that whatever the official line on light-touch regulation, the hands-off posture has limits in practice.
Trump AI Controls Debate Collides With China Theft Accusations
The domestic safety question and the China competition question are now running on parallel tracks, and they are starting to collide. Michael Kratsios, Trump’s senior tech adviser and director of the White House Office of Science and Technology Policy, has accused foreign entities of using ‘tens of thousands of proxies and jailbreaking techniques in coordinated campaigns to systematically extract American breakthroughs,’ according to Nextgov/FCW, which reported on the administration’s April memo.
The specific accusation against Moonshot AI, the Chinese firm behind the Kimi 3 model, is detailed. The Hill reports that Kratsios alleged Moonshot AI built an internal platform specifically to conduct large-scale distillation against US models, switching between multiple access methods to avoid detection, and that the company acquired servers equipped with Nvidia’s GB300 chips, including by accessing them in Thailand, a component the US prohibits from sale to Chinese entities. The Chinese government has consistently rejected these accusations.
The April memo noted, as BBC News reported, that stolen models ‘do not replicate the full performance of the original’ but ‘enable foreign actors to release products that appear to perform comparably on select benchmarks at a fraction of the cost.’ That is the strategic problem in a sentence: China does not need to match American AI to compete with it commercially.
Congress is moving too. The House Foreign Affairs Committee passed the Deterring American AI Model Theft Act of 2026, which would establish a formal government process to deter Chinese AI theft, according to the House Select Committee on the CCP. Treasury Secretary Scott Bessent has warned that Chinese firms could face sanctions for what he called, without much hedging, ‘theft.’
The Federal Communications Commission (FCC) this week added foreign-made humanoid robots and quadruped robots to its ‘covered list,’ meaning they were deemed to pose an unacceptable risk to national security, according to The Hill. The ban also covers foreign-made power inverters used in data centres and renewable energy systems. It does not apply to existing models already authorised for import.
The robotics numbers explain the urgency. Of around 15,000 humanoid robots shipped globally in 2025, Chinese companies Unitree and AGIBOT each shipped more than 5,000, while US counterparts Tesla and Figure AI each shipped a few hundred or fewer, according to technology research group Omdia, as reported by NBC Chicago. China’s foreign ministry responded by accusing the US of overstretching the concept of national security.
Meanwhile, the Pentagon announced dozens more Chinese institutions have been banned from research collaborations with US institutions to prevent unauthorised technology transfer, according to NPR. The administration said it will work with American AI companies to identify distillation activities, build defences, and find ways to punish offenders.
The question now is sequencing. If Trump moves on AI controls before Congress passes the theft deterrence legislation, he risks appearing to regulate his own industry while China runs free. If he waits, the next OpenAI incident may make the delay look reckless. That is the binary he is trying to escape, and so far he has not found the exit.
