Senator Sanders demands AI companies adhere to their own safety commitments, warning of potential Senate intervention #
US Senator Bernie Sanders has told the chief executives of OpenAI, Anthropic, and Meta to 'stand by' their own public promises on artificial intelligence safety. He warned that Congress will step in if they do not development voluntarily.
In a letter dated 10 August and addressed to Sam Altman, Dario Amodei, and Mark Zuckerberg, the Vermont independent said their companies were 'racing ahead' with AI systems that 'nobody can fully understand, predict or control'. He noted each firm had previously pledged to slow down if the technology became too risky.
Sanders Turns AI Giants' Own Safety Pledges Against Them #
Sanders built his letter around commitments the three companies made themselves. He noted that Anthropic said in 2023 it would 'commit to the scaling and/or delay the deployment of new models whenever our scaling ability outstrips our ability to comply with ... safety procedures'.
He also cited Meta's 2025 statement that it would stop development 'if a frontier AI is assessed to have reached the critical risk threshold and cannot be mitigated'. OpenAI had made a similar pledge to 'halt further development' if AI capabilities ever reached a 'critical' threshold.
'That moment is here,' Sanders wrote. 'AI capabilities HAVE reached a critical threshold.'
Sanders Warns of AI-Generated Viruses and Corporate Hacking #
The senator's letter pointed to a string of recent incidents he said showed the technology slipping beyond its creators' grip. He wrote that AI had been 'used for the first time ever to create new viruses', warning the development could lead to bioweapons resulting in 'the deaths of tens of millions of people' if it fell into the wrong hands. The claim has not been independently verified.
Sanders also claimed an OpenAI model had 'hacked into another company's computers', which he called 'a clear violation of federal law'. He said Anthropic and Meta had reported similar incidents involving their own systems after internal reviews. Sanders did not name the company allegedly targeted, and IBTimes UK has not been able to confirm the claim independently.
He quoted Yoshua Bengio, described in the letter as the most cited living scientist in the world, saying the incidents 'should serve as a wake-up call'. Sanders also referenced comments from CIA Director William J. Burns comparing AI models to 'digital nuclear weapons' and 'almost like a doomsday device'.
Senate Ultimatum Self-Regulate or Face Legislation #
Sanders closed the letter with a direct ultimatum. 'Let me be very clear: If you do not take appropriate action now, my colleagues and I in the U.S. Senate will,' he wrote.
The letter lands as AI companies face growing scrutiny over safety failures, even while pouring billions of dollars into ever more powerful models. Should Sanders and other senators move towards legislation, it could reshape how OpenAI, Anthropic, and Meta are allowed to develop and release new systems, with knock-on effects for the millions of people who now rely on these tools daily.
For now, the pressure remains political rather than legal, but the senator's explicit threat marks one of the most direct challenges yet from Congress to the AI industry's self-regulation. Sanders is quoting the companies' own past commitments back at them rather than relying on outside criticism. © Copyright IBTimes 2025. All rights reserved.