When a senior researcher at the Artificial Intelligence company Anthropic warned last week that there was a roughly 10% chance that AI could kill all humans in the next decade, the American public took notice.
Now, the question is whether Congress will take notice — and, more pressingly, whether it will take action.
Over the weekend, Sen. Brian Schatz, D-Hawaii, wrote on social media that he and his colleagues must “move at the pace of the threat, not our own habits and rhythms” to “take swift bipartisan action that meaningfully reduces AI risk.”
But Congress’ cadence of action usually has two settings: slow and not moving at all. And with lawmakers set to leave town to campaign ahead of the midterms, the chances of major legislation moving quickly on Capitol Hill appear slim.
While Schatz dismissed such a downtrodden perspective on Monday — telling reporters the idea that Congress can’t do anything is “insanity” given the stakes — other lawmakers admit it’s a tall order.
Sen. Peter Welch, D-Vt., said it’s going to be “really tough” for Congress to act expeditiously on AI, calling lawmakers “totally awol on this.”
Sen. Bernie Sanders, I-Vt., said he’s hopeful that pressure from constituents back home will drive lawmakers to act, but admitted Congress is, “at best, a slow-moving institution.”
Other lawmakers questioned whether Congress is fully equipped — or remotely willing — to regulate this issue.
Rep. Jared Moskowitz, D-Fla., told MS NOW that Congress has already shown “they don’t have the capacity” to respond to the AI warnings.
And Rep. Tim Burchett, R-Tenn., said he is concerned about the prospect of a “bunch of people up here, literally, who are probably still driving their ‘72 AMC Gremlins” regulating “something they have no idea what it is.”
Of course, there are exceptions to Congress’ usual inertia. In 2008, for instance, lawmakers passed sweeping financial packages — like the Troubled Asset Relief Program, or TARP — to stabilize the financial system.
And in the early stages of the COVID pandemic, Congress passed massive legislation to help families and small businesses navigate the global economic shutdown.
But regulating a sector of the economy before a crisis — rather than throwing money at the problem after the crisis has emerged — has not traditionally been the strong suit of Congress.
Take, for instance, the push to put guardrails around social media, including to protect children — an effort that has languished now for years without resolution.
AI is even more complicated. For one, the technology is evolving at almost incomprehensible speeds. For another, President Donald Trump — whom most Republicans take their cues from — is advocating zero regulation from Congress.
“The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades,” Trump wrote Monday on social media.
Already, there are some lawmakers who seem to agree with the president — or, at least, are echoing arguments to slow congressional action.
On Monday, Rep. Anna Paulina Luna, R-Fla., urged against a “knee-jerk reaction.” Sen. John Hoeven, R-N.D., told reporters that lawmakers “need to understand what we’re dealing with” and make sure they take the “right approach.”
And Rep. Rich McCormick, R-Ga., all but dismissed the idea of government intervention. “It’s your responsibility as a business to do the responsible thing,” he said. “Don’t make a product that destroys the world.”
GOP congressional leaders are making similar defenses for inaction.
On Monday, Speaker Mike Johnson, R-La., warned against moving too fast to regulate AI. “We have to be very careful about this,” Johnson said, adopting an emerging GOP talking point that “we have to maintain our edge over China.”
Johnson said he believes Trump will summon the AI “platform providers” to the White House “this week or early next week” for a meeting.
Senate Majority Leader John Thune, R-S.D., said lawmakers should try to strike a balance. “A light touch, you don’t want to stifle innovation,” Thune said. “But I think you also want to make sure that the more advanced threats can be mitigated.”
Thune is working with some senators — like Amy Klobuchar, D-Minn., and Ted Cruz, R-Texas — on legislation that would require companies to take steps to mitigate catastrophic AI risks. But it’s unlikely anything will come together before the election, or that the legislation will meaningfully address the concerns that senior leaders in the AI sector are expressing.
Meanwhile, Senate Minority Leader Chuck Schumer, D-N.Y., is calling for a classified briefing for senators from the Trump administration about their approach to the technology.
And House Democratic Leader Hakeem Jeffries of New York — who on Monday said the “challenges” presented by the “explosive” growth of AI are not a “hoax” — told MS NOW Congress should “not leave town until something is done decisively to protect the safety and the well-being of the American people.”
(The House is currently slated to leave town to go campaign at the end of the week. The Senate is expected to follow suit shortly thereafter.)
Lawmakers have tossed around ideas on AI for years — and still have not passed anything.
In 2023, Schumer convened a series of closed door, bipartisan sessions for lawmakers — including inviting the CEOs from top tech companies. And by 2024, Schumer and a bipartisan cohort released a “roadmap” for AI policy reforms. But the sweeping framework on AI never materialized and nothing became law.
Some lawmakers, like Sanders, have called for major action, such as a moratorium on new AI data centers. Others have taken a more targeted approach, such as regulating the use of AI in political ads.
A bipartisan duo has introduced “kill switch” legislation that would allow the government to order an AI company to shutdown or slow an AI model that is deemed too dangerous.
Jeffries suggested a Democratic working group is likely to roll out a list of legislative proposals to regulate AI as soon as this week. But the variety of approaches — from calls for sweeping guardrails to Trump’s push for none at all — promises to further complicate an already complicated issue for an institution full of people who never even figured out email.
The pushes for regulation also come at a delicate time: 50 days before Election Day.
While some lawmakers may genuinely want to act, what Congress does or doesn’t do this close to the midterms is inherently political. And neither party is likely to let the other put politics aside to act on AI.
Jeffries, for instance, on Monday accused Republicans of “burying their heads in the sand” on AI.
Still, even if major action is unlikely, Congress could address one of the more immediate concerns with AI: data centers.
This week, the House is set to vote on bipartisan legislation aimed at curbing data center-driven electricity costs for households. The vote is a last-ditch effort for Republicans to address constituent concerns about affordability and AI. But it will do little to address the existential concerns associated with a rapidly advancing technology.
The distinction is important on a policy level — but acting on data centers would be a tangible political win.
A recent NBC News poll found that 69% of Americans oppose the construction of an AI data center in their local area. Seventy percent said they are more worried about AI than excited about it.
The bill introduced by Rep. Gabe Evans, R-Col. — who is defending his seat in a toss-up race — has bipartisan cosponsors and passed out of the House Committee on Energy and Commerce in July with unanimous support. Several of those cosponsors are also in competitive races, like Republican Rep. Ryan MacKenzie of Pennsylvania.
Earlier this month, Mackenzie told reporters that the construction of data centers “very clearly needs to remain a local issue,” but he touted the Ratepayer Protection Act as a way the federal government can intervene.
Mackenzie called the bill “great legislation to make sure that none of those electricity costs are borne by other ratepayers, whether individuals or small businesses.”
On the Senate side, there’s a similar desire to act on data centers.
Sen. Jon Husted, R-Ohio, introduced a Senate companion version of the bill in July, after Husted faced a swarm of political attacks from his Democratic opponent, former Sen. Sherrod Brown, who has spent millions in ads to label him as the face of data centers in Ohio.
But if all Congress does is act on data centers, it would be the latest illustration of lawmakers choosing to regulate an effect rather than a cause.
AI builders are warning that the technology poses an existential threat to the human race. But Congress seems most concerned with — or, at least, most capable of — making sure AI companies pay for the electricity needed to continue making the rapid advancements that researchers are so concerned about.
Mychael Schnell, Syedah Asghar, and Jack Fitzpatrick contributed to this report.
The post AI builders warn the technology threatens humanity. Congress is focused on data centers. appeared first on MS NOW.
