I Tried to Document How Every Member of Congress Feels About AI. The Results Were Strikingly Partisan.

Around the time Rep. Seth Magaziner (D-Rhode Island) arrived in Congress several years ago, he asked outgoing Speaker Nancy Pelosi what she hoped to accomplish after she left leadership. “She said that the most important issue that we were going to face as a Congress was AI, and that she wanted to spend some of her remaining time in Congress really focusing on that,” Magaziner told me recently. “To hear, in 2023, Speaker Pelosi saying that this was the number one issue on her mind, was, I think, when I first realized that this was not just an important issue, but one of the most important.”

Back then, conversations about AI were a rare bipartisan bright spot in an otherwise bitterly divided Congress, and the tech industry, Magaziner recalls, was viewed as relatively helpful. Most significantly, Democrats and Republicans worked together to put forth ideas for regulating AI. In 2024, Republican and Democratic House leaders appointed a bipartisan AI task force — 12 Republicans, 12 Democrats — which, at the end of the year, released a 253-page document containing 85 recommendations. “All 24 of us co-sponsored” those ideas, and “there was a real bipartisan push to do something,” Rep. Don Beyer (D-Virginia) told me recently, noting that “it was a really fun experience.”

Fast forward several years. Interest in AI has exploded in the country as a whole, and there has been no shortage of AI-related efforts in Congress. A few have even passed — most prominently the Take It Down Act, which blocks publication of nonconsensual intimate images, including those generated by AI. And just last week, two House members — a Republican and a Democrat — introduced a bill that would establish some national safety guidelines for AI while also preempting states’ ability to regulate AI developers.

But while the bill is bipartisan in theory, it may face an uphill struggle to garner support on both sides of the aisle — because Congress has changed since it began weighing this topic a few years ago: Views about AI on Capitol Hill now largely track party lines. Democrats, on the whole, want more guardrails, and Republicans mostly favor a hands-off approach. “In just the three years, the brief three years, that I’ve been in Congress, there have been some pretty significant shifts,” Magaziner told me. “A lot of that bipartisanship ground to a halt.”

I discovered the same thing when I recently attempted a comprehensive survey of Congress’ views on AI. I created a database containing public statements on AI or proposed AI-related bills from as many members as possible. My goal was to place members into various camps: generally critical toward AI, generally positive toward AI, or mixed. If a member had sponsored a measure advocating for increased AI safeguards (for example, the GUARDRAILS Act, which seeks to preserve state-level regulations on AI) or had made skeptical-of-AI statements (like this one from Democratic Sen. Ron Wyden of Oregon: “Without new rules with respect to AI you give the executive branch the authority to run roughshod over the privacy rights of Americans”), then I likely classified them as “critical.” If a member had advocated for deregulation to advance AI (for instance, Republican Sen. Mike Lee of Utah wrote an article earlier this year arguing for permitting reform in order to expand AI infrastructure), then I likely classified them as “positive.” Those truly in the center I called “mixed.”

In the end, I found enough material to loosely categorize 256 representatives and senators — 134 Democrats, 118 Republicans and three Independents. I emailed every member to ask if they agreed with where I sorted them. Ten were willing to be interviewed, and 38 more sent back statements. I then adjusted my categorizations accordingly.

Needless to say, my methodology was inherently subjective. For many members, the comments or other data points I found may not fully capture their views. It is also true that very few members of Congress oppose or support AI unconditionally — and that bipartisanship on AI is not entirely dead: A handful of measures have resulted in votes like 99-1 or (in the case of the Take It Down Act) 409-2. And no less a big-tech critic than Sen. Bernie Sanders (I-Vermont) — a sponsor of the AI Data Center Moratorium Act — posted a video last November captioned “Is AI inherently bad? Of course not. The question is: WHO controls it & for what purpose?”

Still, an overall pattern emerged from my database that was hard to dismiss: Nearly 70 percent of Republicans appear generally positive toward AI, while nearly 60 percent of Democrats appear generally critical; and less than 15 percent of Republicans are critical, while less than 15 percent of Democrats are positive. In other words, while the politics of AI among rank-and-file voters are complicated — both Republicans and Democrats mostly oppose local AI data centers, for instance — the situation in Congress increasingly breaks down along party lines. One party, it turns out, is simply far more pro-AI than the other. And the implications of this split could be profound — for legislation, for 2028, for the future of AI itself.

***

The Republicans who responded to my queries reinforced my impression from the database: They see AI as a force for good and, by and large, they want the government to get out of the way. “AI offers the potential for a quantum leap in the quality of life for all mankind,” Rep. Tom McClintock (R-California) wrote in a statement. “The less government is involved in this, the better.”

Another recurring theme in my research was that numerous GOP members are eager for the United States to excel at AI in order to keep pace with China. “This is not simply a technology issue,” Rep. Andy Barr (R-Kentucky) wrote in a statement. “It is an economic competitiveness issue, a national security issue, and a China issue.”

Multiple members confirmed the existence of an AI rift between Republicans and Democrats. “The folks on the other side tend to want a heavier hand in regulation,” Rep. Troy Downing (R-Montana) told me, “and I want a more permissive environment in regulation.”

“The Democrats, unfortunately, are looking to more regulations,” Rep. Vince Fong (R-California) told me. “I think that to do that now will stifle innovation and will prevent America from really gaining and garnering the benefits that AI will provide to our country.”

“A lot of Republicans are focused on the infrastructure side of the equation — how do we build the power generation, transmission lines, pipelines, and data centers needed to support AI?” Sen. Alan Armstrong (R-Oklahoma) wrote in a statement. “Some Democrats are spending more time convincing themselves that AI is driving higher consumer utility bills and that climate change concerns are more important than winning the AI race.”

No Republican I talked to said that there should be no regulation at all. Indeed, the Senate last summer overwhelmingly rejected a proposal originating with House Republicans that would have blocked most state and local AI regulations for 10 years. (The proposal was so widely condemned after it came out of the House that it almost seemed like Senate Republicans had no choice but to reject it.) And some high-profile Republicans, like Sen. Josh Hawley of Missouri, have even gotten media attention for forcefully critiquing AI. But the clear GOP consensus seems to favor a light regulatory touch.

And the philosophical chasm between Republicans and Democrats has only been intensified by President Trump. Though he issued an executive order last week establishing some additional oversight of AI companies, he has generally signaled his opposition to strict AI regulation. “When the president decided to take a hands-off approach at the beginning of his term, that really set things back in Congress,” Magaziner told me.

“If you talk about who is the strongest voice, and the voice to understand why we, America, need to lead in AI technology, it’s the president,” Fong told me. “The president’s leadership is critical.”

***

With Republicans having largely cast their lot with AI, the biggest question might be just how far Democrats will go in the opposite direction.

Some congressional Democrats I spoke to sounded eager to embrace the populist backlash against AI and big tech. “These are not things that you can easily say, ‘Well, I’m pro-business,’” Rep. Chellie Pingree (D-Maine) told me. “If a huge segment of people in your district are losing their jobs, they’re going to be like, ‘Well, what the hell are you doing about me?’”

“Right now, in my opinion, we are not seeing the kind of government national leadership required to protect people … from the extreme concentration of wealth at the top,” Rep. Greg Landsman (D-Ohio) told me. “We’re not protecting people from the cost of these data centers. We’re not protecting people from the cost of a chat bot … saying dangerous things to our kids.”

But my congressional sorting exercise also surfaced complexities on the Democratic side. The initial iteration of my tracker had a dozen more Democrats sorted into the “critical” category, based on publicly available evidence. However, after I emailed members to ask if they agreed with how I sorted them, I immediately began receiving anxious phone calls, emails and texts from communications staffers. The messages were similar: “Critical” is an oversimplification; the senator or representative has a nuanced perspective; they think about AI holistically in a way that fosters innovation and protects the American people; and so on.

Consider Rep. Dave Min (D-California). I had initially categorized him as “critical” based on a strong statement he made at a roundtable earlier this year: “People in our districts across this country are going to start feeling impacts very soon, and if we don’t start thinking properly and aggressively and proactively about the challenges that AI creates, I fear that we’re going to have a revolution on our hands.”

But after I sent him an email, he returned a statement more mixed in tone: “As we are already seeing, AI is unleashing tremendous innovation that is exciting in so many ways. But at the same time, AI creates new challenges that we must be thoughtful and proactive about addressing, including safety concerns and the potentially huge impacts on our workforce.”

I dutifully switched Min and others like him into the “mixed” category. But to me, it illustrated something larger: a clear discomfort among some Democrats in being seen as fully anti-AI. (By contrast, I got only one objection from a Democrat who I had classified as “positive” toward AI. And among Republicans, I got just one note asking for a category switch — a staffer clarifying their boss was very positive toward AI, not mixed.)

All of this led me to wonder whether Democrats will, as a group, pick up the anti-AI mantle in the way that Republicans have more confidently adopted the reverse position. It’s at some level perplexing, given the makeup of the Democratic coalition. AI “can be read as an attack on organized labor, which is a major interest group of the Democratic Party,” says Brian Chen, policy director of Data & Society, a nonprofit research institute that studies the social implications of technology and AI. “Data centers have dramatic ecological impacts. They may well be contributing dramatically to climate change. … The AI sector and big tech writ large have very visibly attached themselves to the Trump administration — showing up at the inauguration — and so optically they make very easy enemies for a Democratic caucus that wants to oppose everything that the Trump administration is doing.”

And yet, there are countervailing factors — chief among them, fundraising. “The AI folks are putting a tremendous amount of money into this election cycle,” Darrell West, a senior fellow in the Center for Technology Innovation at the Brookings Institution, told me. “That obviously affects members of Congress.”

There are other, more principled explanations for Democratic reticence as well. Some Democrats, like Republicans, have concerns about the U.S. falling behind in the AI arms race with China. And most of the Democrats I spoke with emphasized that the goal of AI guardrails is to protect consumers rather than throttle tech companies.

For now, those complicating factors exist uncomfortably alongside a rising tide of anti-AI sentiment among the public — and a sense, at least among some Democrats, that AI needs to be reined in sooner rather than later. Now is not the time for “sitting back and watching and hoping for good,” Landsman told me. “Quite frankly, the longer we wait, the more people get hurt.”

Anusha Mathur is an assistant editor at NOTUS Perspectives.