Congress
Voters are waking up on AI safety. So are these Senate Republicans.
House Republicans have left Washington till November. President Donald Trump is decrying concerns about the issue as a “HOAX.” But a few Senate Republicans are trying to nudge their party into doing something — anything — about artificial intelligence.
There isn’t much binding them together aside from the “R” next to their names, but Sens. John Curtis, Josh Hawley and John Kennedy have each been closely attuned as voters suddenly grow alarmed — not only about the data centers powering the AI industry, but the recent insider warnings about potentially catastrophic consequences for humanity.
Curtis is a Utah moderate who has been among the few in his party to take seriously scientists’ warnings about climate change. Hawley is a populist Missouri conservative who has long been skeptical of Silicon Valley business titans. And Kennedy’s simple-Louisiana-country-lawyer persona belies a relatively sophisticated understanding of frontier AI developments.
The trio is not coordinating efforts, but their individual actions were part of a clear push from the GOP’s back benches over the past week to get party leaders to wake up and address an issue they believe could tilt the upcoming midterms and certainly the presidential and congressional races in 2028.
“We might want to be on the side of voters — just a thought,” Hawley said, noting that while he has his own “views” on how Washington should respond, party leaders have offered members no guidance on how to approach the topic.
Hawley is taking matters in his own hands this week by convening a hearing in the Judiciary subcommittee he chairs on AI-powered Flock cameras. He’s also using another subcommittee gavel to investigate OpenAI over the runaway model that saw hundreds of rogue agents escape testing and hack another tech company.
Both Curtis and Hawley have teamed up with Democrats, who have generally been more aggressive in calling for AI safety measures — particularly since a former OpenAI and Anthropic employee went public saying advanced models could “could kill us all by the end of the decade.”
Hawley and Sen. Richard Blumenthal (D-Conn.) — who both sit on the Judiciary and Homeland Security and Governmental Affairs committee — pushed congressional leaders last week to hold a vote on their “risk evaluation” legislation, arguing that “frontier AI demands frontier legislation.”
Curtis, meanwhile, joined with a fellow Commerce Committee member, Sen. Lisa Blunt Rochester (D-Del.), to call for “immediate public hearings” to “work through solutions that maintain America’s competitive edge in development while ensuring that these technologies serve human interests and remain fully under human control.”
Curtis said in an interview lawmakers need to “show that we’re adults in the room, that we can talk about this, that we can find solutions.”
“I don’t want to say we’re making it too hard, but I think we’re making it too hard,” Curtis said. “And what I mean by that is, there’s a lot of false narratives out there — we need to just jump on this immediately and clamp it down, or we need to not touch it, and all of that only I think causes anxiety back home.”
Both efforts could be seen as attempts to jolt more senior senators who are in positions to advance AI-related legislation. So far they are treading carefully amid skepticism among Senate Republicans that they will ever be able to overcome Trump’s opposition on AI regulations and recalcitrance among House GOP leaders, according to three people granted anonymity to describe private conversations.
They include Judiciary Chair Chuck Grassley (R-Iowa), who indicated in an interview that he had no plans to hold hearings or take other action related to AI.
“I think that I’d be getting into the jurisdiction of the Commerce Committee,” Grassley said, directing further questions to Commerce Chair Ted Cruz (R-Texas).
Cruz has been more deeply engaged than Grassley in AI matters, but his plans for action remain uncertain. He said last week he is still hoping to hold a committee vote on AI safety legislation this month before the Senate leaves for the final stretch of pre-midterm campaigning.
Cruz is engaged in negotiating a revised version of an AI safety bill that Senate Majority Leader John Thune and Sen. Amy Klobuchar (D-Minn.) first introduced in 2023. But multiple Commerce Committee Republicans said they have so far heard zilch behind the scenes about what the committee could do, and some doubt they will be able to get an agreement in time to act before the Senate leaves — possibly as soon as the end of this week.
While Thune himself has been the most engaged on AI issues among top GOP leaders, he still advocated for a “light touch” on industry regulation last week while acknowledging the “considerable amount of interest” among lawmakers in taking action.
Sen. John Barrasso of Wyoming, the No. 2 Senate Republican, said — as Hawley suggested — that he isn’t giving members guidance on how to talk about AI.
“People have opinions and they’re going to make their opinions known and talk to their voters,” he said.
And then there’s Kennedy, who recently said the industry is run by “high-IQ stupid people” and was the rare Republican who attended a briefing put together by Sen. Bernie Sanders (I-Vt.) last week to meet with AI experts to talk about potential safeguards for the technology.
He took his AI concerns straight to the Senate floor Wednesday, seeking to pass legislation this week that would require companies to install so-called “kill switches” for their AI models. In a speech before seeking unanimous consent to pass the bill, he delivered a homespun but technically detailed summation of the OpenAI swarm attack and the risks posed by “recursive self-improvement.”
His bill was blocked by Sen. Rand Paul (R-Ky.), who instead proposed creating a panel to make recommendations about potential guardrails.
Kennedy said appointing a committee would be “the weenie way out” after warning his colleagues from the Senate floor that Americans were growing “very skeptical” of AI and the tech titans behind it.
“They’ve seen this vampire movie before,” he said.
In a separate interview, Kennedy said he hadn’t heard “anything” about action in the Senate to address AI and that lawmakers who ignore the growing voter uproar around the issue were courting peril.
“We really need to send a memo around explaining we’ve got midterm elections coming,” he said. “I may send that out.”