Congress
Inside John Thune’s AI dilemma
John Thune has urged Congress to act on artificial intelligence for years. Now, as concerns about the rapidly advancing technology explode, the Senate majority leader is moving carefully.
A former Commerce Committee chair with a long-running interest in tech policy, Thune is locked in negotiations over a bipartisan AI safety bill. The measure, which he introduced a version of in 2023, would establish some safety and security guardrails for the riskiest models — giving the South Dakota Republican, on paper at least, a solution to Washington’s growing AI nightmare.
But the reality is much messier. Despite growing alarm bells from industry insiders about the potentially catastrophic fallout of runaway AI, Republicans and Democrats remain far apart on how stringent a regulatory framework should be — particularly if it overrides state safety laws.
“It’s probably TBD” if an AI bill could pass by the end of the year, Thune told reporters Monday as global players took stock of a host of warnings from AI lab insiders — as well as brushoffs from President Donald Trump and tech-industry investors.
“There are going to continue to be a lot of conversations about potential paths forward, but getting anything done in the near term is going to be challenging given the other stuff we’re dealing with,” he added.
In addition to negotiating his own bill, Thune will need to navigate Trump’s growing recalcitrance. The president dismissed the idea that unregulated AI models could threaten humanity as a “HOAX” Monday, even as his party faces backlash from voters who have grown increasingly skeptical of the underlying technology and the companies developing it.
Thune is also facing competing factions within his own party ranks, not to mention among Democrats, about how heavy of a hand Congress should have in determining AI guardrails. Across the Capitol, Speaker Mike Johnson and other House GOP leaders are warning against moving too quickly on potential regulations, while some other traditional small-government conservatives are suggesting Congress shouldn’t stay on the sidelines.
Senate Republicans are expected to discuss AI policy during a weekly closed-door lunch Tuesday — their first in-person meeting since an industry researcher warned last week that AI could destroy humanity and some company leaders called over the weekend for a joint slowdown on model development.
“We’ve been having meetings about AI for several years, trying to figure out the best thing to do about it,” said Sen. John Barrasso of Wyoming, the No. 2 party leader. “We’re going to continue talking about it, probably more aggressively now than we had.”
Historically, Republicans have expressed concern about overregulating the AI industry out of fear the government could stifle American innovation and, ultimately, cede ground to China. Thune appeared to flick at that concern within his party Monday, saying he believed there was a role for Congress but that lawmakers should take a “light touch.”
“There is a way in which Congress can put guardrails around those types of more consequential threats, without in any way harming our ability to stay ahead in the AI race, which I think is also very important,” Thune said, arguing that the bill he is negotiating with Sen. Amy Klobuchar (D-Minn.) would strike that balance.
But others on Capitol Hill engaged on AI matters who distrust the industry are increasingly convinced a heavy hand is needed given the risks exposed in recent runaway-model incidents.
Sen. Maria Cantwell of Washington, the top Democrat on Senate Commerce, in an interview Monday criticized Republicans over the weekend for dragging their feet and mocked the GOP’s past approach to technology.
She recalled experts on a bipartisan panel warning lawmakers years ago “this could go rogue, so you need to do something, but the Republicans kept saying, ‘No, no — light touch, light touch, light touch, light touch.’”
“I’m pretty sure they’re going to still say that,” Cantwell added.
When told that Thune did in fact utter the phrase “light touch” Monday, she replied, “Okay, that’s bad.”
It’s unclear if Thune’s former Democratic partners on AI regulation will lock arms with him this time around.
Klobuchar has a yearslong history of teaming up with Thune on a variety of issues — including the 2023 version of their AI legislation. But Thune said Monday he spoke with Klobuchar about a revised bill over the weekend and said she was “reluctant to move forward” without more Democratic support.
In a brief interview Monday, Klobuchar said she’s working to strengthen the proposal: “We’re continuing to negotiate.”
In a subsequent statement responding to Thune’s comments, she said it is “clear” from talks that there is bipartisan agreement “that something needs to be done.”
“I’m committed to continuing to work with Senators Thune, Cantwell, and [Commerce Chair Ted] Cruz to ensure we have a bill with strong safeguards,” Klobuchar continued. “I do not expect the chair and ranking member to agree with every provision in the bill but hope we can reach consensus. Time is of the essence.”
The House is expected to leave town this week until after the November elections, and the Senate could follow suit as soon as next week. Neither expected to take up sweeping AI safety legislation before that. A separate bill narrowly focused on the energy impacts of data center development could make it to Trump’s desk before then, but only if all 100 senators agree to expedite it.
Cruz, who has done little on the subject of AI safety since he became chair last year, admitted in a recent interview with POLITICO that the threats being raised inside the top AI labs are “scary as hell.”
But the Texas Republican hasn’t yet scheduled a long-sought committee vote on AI bills after multiple delays. Asked Monday about progress on the Thune-Klobuchar bill, Cruz said, “We’re working on it. We’re not there yet.”
Consensus has also been tough to find inside the Democratic caucus. Progressives including Sen. Bernie Sanders (I-Vt.) have called for an immediate moratorium on the development of AI superintelligence, while others have proposed a more moderate approach.
Sen. Brian Schatz (D-Hawaii), a Commerce member, said in an interview that lawmakers don’t need to resolve every debate over the future of AI — they just need to establish a baseline standard for transparency and oversight of the industry.
“What we need to do is take down the risk of these models either jumping out of the sandbox or being released into the wild without any third party certifying to their safety,” he said. “And if we could just do that and then fight about the rest of it in a few weeks or months, that would be progress.”
Jennifer Scholtes contributed to this report.