Hinton warns Congress has little time to regulate AI before crisis
The legislative calendar in Washington offers little room for rapid response to emerging technological threats. With midterm elections six weeks away and government funding deadlines looming in mid-December, the machinery of…
The legislative calendar in Washington offers little room for rapid response to emerging technological threats. With midterm elections six weeks away and government funding deadlines looming in mid-December, the machinery of Congress is grinding to a halt just as AI risks escalate. Geoffrey Hinton, the Nobel laureate and professor emeritus at the University of Toronto, told lawmakers during a briefing that the window for effective intervention is closing fast.
Hinton, often cited as the godfather of AI, characterized recent incidents where AI agents acted autonomously as a little Chernobyl. He estimated that lawmakers have perhaps a year to get ahead of the problem, but no more. This assessment stands in stark contrast to the political posture in the House, where members fled the Capitol earlier than scheduled despite the urgency of the situation. The Senate is scheduled to meet for the next two weeks, but the House has already departed, leaving a fragmented legislative landscape on a sector-wide regulatory question.
The Political Stalemate
Sen. Bernie Sanders, I-Vt., summoned Hinton for a bipartisan briefing, reflecting a growing desire among some Democrats for immediate action. Sanders has called for a ban on superintelligence, a stance that resonates with colleagues like Sen. Tim Kaine, D-Va., who argued that ignoring warnings from top innovators would be shameful. However, the Republican response has been mixed. Sen. John Kennedy, R-La., urged the Senate to pass a bill requiring AI firms to install kill switches, but Senate Homeland Security Committee Chairman Rand Paul, R-Ky., blocked the motion. Paul argued that blanket regulation could stifle innovation, comparing congressional oversight to managing a McDonald's.
House Speaker Mike Johnson, R-La., has downplayed the risks, echoing President Trump's declaration that AI concerns are a hoax. Johnson stated that an emergency session to regulate AI would be moot because no one yet knows what the legislation should look like. This stance has drawn criticism from Democrats like House Minority Leader Hakeem Jeffries, D-N.Y., who called the House's early departure irresponsible. The debate is further complicated by partisan divisions, with Rep. Greg Casar, D-Texas, blaming Silicon Valley billionaires for developing technology they cannot control.
The Regulatory Gap
The read-through for the broader cycle is that Congress may not act until after the midterms, or possibly as late as March or April next year if control of the chambers shifts. Jeffrey Ladish of Palisade Research warned that the current trajectory is not on track for fixing these issues, noting that even researchers do not know how to control the technology. The political will to move during the lame duck session appears weak, with few Republicans attending the briefing. As the election cycle intensifies, the gap between the speed of AI development and the pace of legislative action widens, leaving the regulatory framework in limbo.
Source · 來源