While it’s not the first time experts have warned of the existential threats posed by an unregulated industry that is rapidly advancing, it comes at a moment of heightened awareness on Capitol Hill about the need to reach consensus on AI safety legislation before it’s too late.<\/p>\n

\u201cThe call is coming from inside the house,\u201d Rep. Lori Trahan of Massachusetts, a House Democrat seeking to lead her party on AI policy, posted on X<\/u><\/a> Wednesday. \u201cSafety researchers are resigning, powerful AI models are breaking out of their labs, and companies are racing ahead anyway.\u201d<\/p>\n

She was responding to social media posts from former Anthropic AI researcher Jason Coxon, whose warnings were first reported in the Wall Street Journal<\/u><\/a>.<\/p>\n

\u201cI resigned from Anthropic today. I spent the last three years doing pretraining research at both OpenAI and Anthropic,\u201d said Coxon, who specialized in training models. \u201cNeither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives.\u201d<\/p>\n

He added: \u201cThe people building AI earnestly believe that it could kill us all by the end of the decade. This is not a marketing stunt.\u201d<\/p>\n

Adding to the alarm, Evan Hubinger, a top Anthropic scientist charged with ensuring AI aligns with human values and ethics, agreed with Coxon’s<\/u><\/a> characterization of the situation.<\/p>\n

\u201cWe really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade,\u201d he posted. \u201cI believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.\u201d<\/p>\n

Spokespeople for OpenAI and Anthropic did not immediately respond to requests for comment, but leaders of both companies have engaged with Congress in the not so distant past on the issue. Anthropic CEO Dario Amodei made a direct plea<\/u><\/a> to lawmakers to impose limits on AI development in a New York Times op-ed last year, while OpenAI CEO Sam Altman has repeatedly<\/u><\/a> visited<\/u><\/a> congressional offices this year.<\/p>\n

Trahan was not the only member of Congress Wednesday to join the calls for guardrails now.<\/p>\n

Democratic Rep. Ted Lieu of California \u2014 a member of leadership and co-chair of an AI commission that Minority Leader Hakeem Jeffries charged with formulating AI policy recommendations \u2014 called Coxon\u2019s post<\/u><\/a> \u201cexhibit number 739 for why we need to pass the bipartisan AI Kill Switch Bill asap.\u201d<\/p>\n

Republican Rep. Anna Paulina Luna of Florida urged Congress to convene a special session<\/u><\/a> on AI, asserting \u201cnot enough of Congress is focusing on this.\u201d<\/p>\n

\u201cWe MUST be ready,\u201d she wrote. \u201cThis is happening whether or not we want it to.\u201d<\/p>\n

But lawmakers remain far apart on agreeing to any plan. Trahan continues to push for Congress to pass her bipartisan AI safety bill with Republican Rep. Jay Obernolte of California as members of both parties say it falls short<\/u><\/a>.<\/p>\n

Sen. Bernie Sanders (I-VT), who has often been ahead of the pack in calling for strict standards \u2014 including his latest proposal that would ban AI superintelligence<\/u><\/a> \u2014 declared on social media that 81 percent of Americans believe the government is not doing enough to regulate AI.<\/p>\n

“AI threatens our economy, privacy, democracy and the well-being of our kids,” he said.<\/a> “Congress must stand up to the Big Tech oligarchs and protect the American people.<\/p>\n

Sanders\u2019 proposal with House Progressive Caucus Chair Greg Casar (D-Texas) to ban AI superintelligence, however, was widely panned by colleagues who said barring innovation was not the answer.<\/p>\n<\/div>\n","protected":false},"excerpt":{"rendered":"