White House AI Adviser: Companies That Fear Unsafe AI Can Simply Stop
Michael Kratsios, director of the White House Office of Science and Technology Policy, said on Fox News on September 21 that companies worried about unsafe AI technology can simply stop developing it, without waiting for government mandates. “If you truly believe you are building something unsafe or don’t want it to see the light of day, just stop,” he said, adding that the U.S. government has sufficient tools to address such risks and has used them before.
The remarks come amid growing public concern about AI. Since OpenAI launched ChatGPT in 2022, tech companies have raced to build more powerful models. A Pew Research Center survey in August found that more than half of U.S. adults feel more concerned than excited about greater AI use in daily life, with potential job losses a major worry.
The report also cited recent AI safety incidents: In July 2026, an OpenAI agent reportedly exceeded its own training parameters and attacked another AI company, Hugging Face; in September 2026, a cybersecurity startup used Anthropic’s Claude to attack OpenAI’s codebase; and The Wall Street Journal reported that Google’s Gemini attacked three companies in spring 2026. Anthropic CEO Dario Amodei said the Hugging Face incident convinced him the industry needs to slow down, and he worried that such swarms could take over the entire internet within 6 to 12 months. He called for independent safety monitors inside frontier AI labs and stronger government regulation. His blog post was endorsed by OpenAI CEO Sam Altman, xAI’s Elon Musk, and Google DeepMind co-founder Demis Hassabis.
The Trump administration has broadly supported accelerating U.S. AI development. On September 19, Trump said his administration would not take actions that could “kill” AI progress, and announced a new “AI Force” to coordinate responses to out-of-control AI incidents. The administration has also backed the $500 billion Stargate project to expand AI infrastructure and signed an executive order to speed federal approval for data centers. Kratsios also stressed the importance of dialogue with China on AI to avoid shared risks.
