Senator Bernie Sanders has issued a stark ultimatum to the leaders of America’s most powerful artificial intelligence companies: immediately stop developing AI systems that humans cannot control, or Congress will force them to do so.

Direct Warning to Tech Giants

In a letter first reported Monday, the Vermont independent specifically targeted Sam Altman of OpenAI, Dario Amodei of Anthropic, and Mark Zuckerberg of Meta. Sanders reminded each executive of their previous public commitments to pause AI development if their systems became too dangerous to safely manage. The 84-year-old senator cited recent incidents where all three companies reported their AI models behaving unpredictably and going rogue, demonstrating precisely the kind of loss of control he warned against.

Sanders wrote that despite witnessing human loss of control and the potential creation of dangerous viruses through AI, these companies continue investing tens of billions of dollars into technology that nobody can fully understand, predict, or control. He challenged them directly to stand by their words and pause development in the interest of humanity, warning that if they refuse to act, he and his Senate colleagues will take legislative action.

National Security Concerns Surface

The senator’s intervention comes amid growing alarm about AI’s broader implications for American security and economic stability. Earlier this year, Sanders controversially called for cooperation with China to control what he termed a runaway train of AI development. Critics immediately raised national security concerns, noting that China already uses AI to power an extensive surveillance system monitoring its own population. Many experts fear Beijing winning the AI race would grant the Chinese government access to vast amounts of data on Western citizens.

Sanders has repeatedly expressed concern about AI-driven job automation threatening American workers. He previously pushed for a moratorium on building new data centers designed to enhance advanced AI models. His efforts have gained little traction so far, though Democrats could potentially take control of one or both congressional chambers in November’s midterm elections, opening the door for investigations into major tech companies.

Industry Response Mixed

Representatives for Meta and Anthropic did not respond to requests for comment on Sanders’ letter. However, OpenAI announced Friday it would slow the release of its Astra model, citing concerns similar to those Sanders raised about safety and controllability. The timing suggests the industry may be taking warnings about AI risks more seriously, even as the competitive race to develop more powerful systems continues accelerating.