AI insiders say their technology could doom humanity. Will Congress act?

Por Riley Rogerson and Kelsey Brugger — POLITICO – TOP Stories

Members of Congress are confronting a grim reality: They are ill-equipped to rein in advanced artificial intelligence that one expert warns could “kill us all” by the end of the decade.

A viral social media post this week from ex-Anthropic researcher Jacob Coxon cautioned that AI companies are “gambling with our lives” as they race to develop superintelligent models — prompting alarm from lawmakers on both sides of the aisle and new calls to action.

In conversations with POLITICO throughout the day Wednesday, more than a dozen lawmakers and aides from both parties said they were hopeful Congress would start taking the issue seriously. By Wednesday evening, OpenAI’s chief global affairs officer, Chris Lehane, was pledging to “work with Congress” to implement “mandatory, capability-based national AI safety regulation.”

But many also conceded that Congress is currently sleepwalking into a crisis, far from ready to tackle the stakes of the moment in light of both philosophical differences and intractable partisan politicking.

“I think we’re finally hitting a tipping point,” Rep. Don Beyer (D-Va.), who co-chairs the House AI Caucus, said in an interview. “I’m just so grateful that the whole world is finally paying attention … and now the question is, what do we do?”

Mounting worries inside the tech industry about a looming AI apocalypse are beginning to resonate with more members of Congress. Building on the panic seven weeks ago when hundreds of OpenAI-generated “agents” went rogue and hacked another AI platform, Hugging Face, lawmakers Wednesday called for radical action in the immediate aftermath of Coxon’s X posts late Tuesday night.

“There’s broad consensus — and in speaking with the AI companies directly, they want a regulatory framework,” Rep. Mike Lawler (R-N.Y.), a vulnerable moderate incumbent who has introduced bipartisan AI safety legislation, said Wednesday.

He called it “one of the biggest issues that Congress has to tackle over the next two years.”

It’s far from certain this desire to make progress on the issue will translate into meaningful legislative activity — and certainly not before the end of the year, with some lawmakers questioning whether their colleagues take the matter seriously or fully understand how the technology even works.

Democratic Rep. Sam Liccardo of California, who leads AI policy for the centrist New Democrat Coalition, said in an interview his team recently struggled to convince other members to attend a series of events at the Library of Congress that featured OpenAI CEO Sam Altman, Anthropic CEO Dario Amodei and Nvidia CEO Jensen Huang.

“There have been serious conversations among a small number of us, and we need that number to be larger. There is a percentage of my colleagues who are still trying to spell AI,” Liccardo said, adding, “we need to dramatically increase the base knowledge of Congress if we’re going to legislate meaningfully. And I know that is not going to happen overnight.”

Rep. Nathaniel Moran (R-Texas), who introduced a bill this summer to create an AI “kill switch” with Rep. Ted Lieu (D-Calif.), said in an interview he thought efforts to put checks on advanced models would be a priority in the next Congress. Quicker action, he agreed, would be a heavy lift.

“The number of days that we have left between the election and the end of the year is going to be very tough to educate members and then move a meaningful bill,” he said. “The broad scope of members on both sides of the aisle are trying to get caught up.”

There’s also not yet consensus on the appropriate reaction and policy response to the threat posed by superintelligence. No committee in either chamber has considered a substantive proposal to put guardrails on AI development amid disagreements over how aggressive the federal government should be in regulating the private sector and stifling innovation.

Senate Commerce Chair Ted Cruz (R-Texas), who convened a markup of AI-adjacent bills in August that steered clear of addressing safety regulations, said in an appearance on ABC’s “The View” Wednesday that the new warnings were “highly concerning” and endorsed passing legislation to prevent “catastrophic risk.”

He pointed to a pending bipartisan bill from Senate Majority Leader John Thune and Sen. Amy Klobuchar (D-Minn.) that is expected to represent a more targeted approach than a version of the legislation that advanced out of the Commerce Committee in the previous Congress.

Klobuchar and Thune, however, have provided no timetable for when their new measure might be formally introduced. Thune’s office did not respond to a request for comment, while Klobuchar in a statement said only that she would “continue to work toward a bipartisan agreement on legislation for commonsense guardrails to address the risks posed by AI systems,” while adding, “it’s clear we need to act now and not wait.”

Across the Capitol, House Democrats made especially clear in the aftermath of Coxon’s X posts that they will seek to move quickly if they win the majority in November, with some now discussing whether to establish a select committee on AI.

House Minority Leader Hakeem Jeffries convened a Democrat-only AI commission this term to formulate policy recommendations after Speaker Mike Johnson declined to reconstitute a bipartisan AI taskforce, and co-chair Josh Gottheimer (D-N.J.) projected confidence Wednesday the group’s forthcoming report would pave the way for meaningful legislative action on AI safety.

Passing AI safety legislation will require buy-in from Republicans, however — and Gottheimer, asked if Congress is moving fast enough, blamed GOP leaders for lack of momentum.

“What you’re really asking me is if the Republican leadership is working fast enough on this, right? Because they’re the ones who control the agenda,” he said. “I think you’ve seen a lot of action out of members of Congress. I think our commission has been incredibly active.”

Other Democrats joined Gottheimer in blaming Republicans for slow-walking legislation.

“It only would take one phone call from the speaker to say, ‘I want five Democrats and five Republicans in a room, and I want them to emerge from that room next week with legislation that we can bring to committee and then bring to the floor,’” said Rep. Greg Landsman (D-Ohio), a member of the House Energy and Commerce Committee. “That’s it.”

Spokespeople for Johnson and Jeffries did not return requests for comment.

Even if Democrats take back gavels after the midterms, they, too, lack a clear plan to address the growing threat of unchecked AI. Progressive leaders, including Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas), are advocating for a ban on superintelligence — a proposal that more moderate Democrats claim has no viable path to passage.

“Bernie Sanders,” said Gottheimer, “wants to put it in the box and throw it away, but that’s not living in reality. I think we need a reasonable, smart, commonsense approach that’s realistic but also sets up mandatory rules of the road — legislative rules of the road.”

Rep. Lori Trahan (D-Mass.) brokered a bipartisan safety bill with Rep. Jay Obernolte (R-Calif.) that the two lawmakers have pitched as a consensus proposal, but it has so far failed to win the support to advance.

“You cannot just put your head in the sand on AI and say, ‘I hope it works out.’ You can’t just go in and say, ‘Government’s got the solution — they’re just going to ban it all.’ That’s just not going to work,’” Rep. Chip Roy (R-Texas), a staunch conservative who is leaving Congress at the end of the year, said in an interview. “We’re going to have to be thoughtful as leaders and actually have some friggin’ leadership.”

Owen Dahlkamp, Gabby Miller, Katherine Long, Mia McCarthy and Alex Gangitano contributed to this report.

Fonte: POLITICO – TOP Stories