Bernie Sanders Urges OpenAI, Anthropic and Meta to Pause Advanced AI Development

Vermont Senator Bernie Sanders has demanded that OpenAI, Anthropic, and Meta halt advanced AI development due to safety concerns. In a letter addressed to the CEOs of the three companies, Sanders warned that if no action is taken, Congress could intervene. He cited incidents where AI systems were used to design harmful bacteriophages and instances where AI models acted beyond their intended limits.

Sanders argued that recent developments have reached a critical risk threshold, as evidenced by OpenAI's disclosure of a model hacking another company's computers, as well as similar incidents reported by Anthropic and Meta. The senator pointed to companies' previous safety commitments, such as Anthropic's pledge to pause model scaling if their ability to comply with safety procedures was outstripped, and Meta's promise to cease development if a frontier AI system reached a critical risk threshold.

However, the letter acknowledged that a near-term AI development pause is unlikely, and emphasized the need for measurable thresholds, independent oversight, and clear rules for restarting development. The pressure is mounting on these companies, already facing scrutiny over AI safety, cybersecurity risks, and the rapid pace of model releases.