Progressives push legislation to halt development of human-surpassing AI
Sen. Bernie Sanders and Rep. Greg Casar introduced a bill on Wednesday that would prohibit AI systems exceeding human cognition across all domains, as well as pause advanced AI work until a new federal agency issues safety rules. The lawmakers argue that tech leaders themselves admit losing control, making further advancement irresponsible and potentially catastrophic. The proposal comes amid a week of AI discussions tied to Chinese President Xi Jinping's U.S. visit and President Trump's calls for American dominance in the field.
The proposal targets AI that surpasses human reasoning broadly or poses existential risk, halting such work until a new federal department issues safety rules. It arrives during President Trump's U.N. remarks, where he reframed the technology as "superintelligence" and stressed a competitive race against China, contrasting with the lawmakers' push for international limits.
The bill would require firms to submit pre-development plans and obtain approval before deployment, while banning "precursor" traits like enhancing weapons design or self-modification. Casar acknowledged industry resistance to a domestic pause without global agreements, arguing voluntary risk assessments are insufficient. Rep. Khanna separately urged Trump and Xi to focus on AI safety.
This legislation could reshape AI development timelines, potentially slowing commercial deployment while creating a new federal oversight body. If enacted, it may affect tech firms' investment strategies and researchers' priorities, while also influencing public trust in AI governance. The emphasis on international agreements could either foster global cooperation or deepen geopolitical tensions, depending on other nations' responses. Society may benefit from reduced existential risk, but could also face slower innovation in beneficial applications like medical research.