Lawmakers Seek Unified AI Oversight Amid Safety Warnings

More than twenty members of the US Congress have begun backing a push for stricter artificial intelligence regulations this week. This legislative shift follows public warnings from a researcher who left Anthropic, citing fears that AI companies are taking dangerous shortcuts with public safety.

The momentum comes as several competing bills sit in the legislative queue, ranging from mandatory safety shutdowns to temporary pauses on advanced development. While the political will appears to be growing, the path to passing any single law remains complicated by jurisdictional overlaps and the upcoming election cycle.

Researcher resignation sparks political urgency

Jacob Coxon, a former OpenAI employee who recently worked at Anthropic, announced his departure in a social media post that reached a massive audience. He warned that the people building these systems believe they could cause fatal harm within the next decade. His comments prompted a response from Anthropic’s alignment science lead, who personally estimated the risk of such an outcome to be above ten percent.

Rep. Lori Trahan, a Democrat from Massachusetts, stated that her phone has been ringing with calls from colleagues on both sides of the aisle. She described the situation as reaching a tipping point, with rank-and-file members demanding concrete action rather than further debate. This cross-party alignment suggests that safety concerns are becoming a central issue in congressional discussions.

Calls for a dedicated oversight committee

Senator Ruben Gallego of Arizona has written to Senate leadership, urging the creation of a bipartisan select committee on artificial intelligence. He argued that current oversight is fragmented across multiple committees, leaving no single body with a complete view of the technology’s impact. He proposed that this new committee be granted subpoena power to gather evidence and produce actionable legislative recommendations.

Rep. Nathaniel Moran, a Republican from Texas, echoed these sentiments, noting that innovation and safety do not have to be mutually exclusive. He emphasized the need for deliberate policymaking that addresses risks without stifling progress. This stance reflects a broader consensus that the current patchwork of regulations is insufficient for managing the rapid pace of AI development.

Competing bills face legislative hurdles

Several pieces of legislation are already in motion, though none have advanced significantly. The FRONTIER Act aims to create governing rules for how advanced models are deployed, while the AI Kill Switch Act would require companies to maintain functional mechanisms for halting their systems. A third proposal, the Ban Artificial Superintelligence Act, suggests a temporary pause on development until federal safety rules are established.

According to reporting from GN technics/ai (en-US), the prospects for any of these bills becoming law in the near term are slim. Both chambers of Congress are on recess ahead of the midterm elections, and President Donald Trump has recently dismissed extinction-level concerns, prioritizing economic competitiveness instead. Meanwhile, industry leaders like Chris Lehane from OpenAI are calling for mandatory, capability-based national regulation, acknowledging that the policy window for sensible oversight is currently open.