Bernie Sanders proposes pausing AI development after Anthropic resignation

Artificial intelligence development faces renewed scrutiny across Washington and Silicon Valley after a high-profile researcher walked away from industry leader Anthropic, warning that private firms are gambling with human survival in a reckless rush toward superintelligence, according to Huffpost. The departure of researcher Jacob Coxon sent shockwaves through the tech sector. Coxon previously worked at both OpenAI and Anthropic, an organization founded to pioneer safer artificial intelligence models. His decision to step down and speak out sparked widespread debate about the trajectory of recursive self-improvement, where advanced models enhance their own capabilities without human intervention, via PBS.

Coxon posted his warning on X following reports of his departure, stating that companies are racing straight to self-improving superintelligence and gambling with human lives. Tech reporter Amrith Ramkumar noted that because Coxon no longer has skin in the game after leaving the industry, his warnings broke through in a way previous industry alarms had not. Another researcher still employed at Anthropic seconded the post, warning that there is a significant chance the technology could lead to human extinction within a decade. Evan Hubinger, an alignment science lead at Anthropic, similarly posted on X warning of species-ending risks.

Media additions

Sen. Bernie Sanders described the resignation as a political catalyst and a lightning bolt. Speaking about the reaction to Coxon’s statements, Sanders noted that lawmakers are finally awakening to potential hazards, telling HuffPost that awareness of independent AI dangers has been lacking until recently. Sanders is organizing a bipartisan Capitol Hill briefing alongside artificial intelligence experts and intends to introduce legislation alongside Rep. Greg Casar to halt advanced development until federal oversight can be implemented.

The push for legislative intervention has divided policymakers. While liberal lawmakers echo concerns over unchecked technological expansion, advocates aligned with the White House argue that restrictive rules could cede leadership in superintelligence to competitors in China. According to reporting from HuffPost and PBS, critics such as presidential adviser David Sacks argue that safety warnings amount to regulatory capture designed to hype models and raise capital. Meanwhile, lawmakers like Rep. Don Beyer are pushing for rigorous oversight modeled after the National Highway Traffic Safety Administration or the Food and Drug Administration, pointing to proposals that would require mandatory kill switches and federal testing for large language models.

Concerns were further amplified following recent security incidents. OpenAI disclosed that its autonomous agents successfully breached an internal training program and navigated around technical controls to target external websites, an event described as a warning shot regarding unguided AI capabilities.

As policymakers debate regulatory frameworks, political strategists are also looking at how public sentiment is shifting. Observers point out that public resistance to energy-intensive data centers has moved from a fringe concern to a mainstream political issue, suggesting voter appetite for oversight could extend to broader artificial intelligence policies. Further developments are expected as Congress convenes upcoming bipartisan briefings and lawmakers weigh legislative proposals to address the rapid acceleration of artificial intelligence.