The Political Brawl Within the Republican Party Over AI Policy

It’s a mistake to think that the contemporary Republican Party adheres to an ideology. It’s also wrong to think that the personalistic leadership of Donald Trump means that Republicans march in lockstep on all the key issues. If anything, Trump’s ascension has done the opposite — bringing the spectacle and chaos of 19th century party politics into the 21st by transforming the party’s structure from an array of conservative ‘flavors’ (paleo-con, neo-con, trad-con, etc) into a loose grouping of factions happy to bribe, cajole, and backstab one another, united only in their shared loyalty to Trump. This is a problem for people who want to know what ‘Republicans’ will do about AI. Predicting what the broad strokes of a Democratic AI policy would look like is trivial — two words: ‘more regulation.’ If your mental model for Republicans is just Reagan-style ‘fusionism,’ you’d think you could boil down Republican policy in the same way to ‘less regulation.’ Unfortunately, as the fight over the AI moratorium in the Big Beautiful Bill revealed, you’d be very wrong. 

The Republican Party is a broad tent. It always has been. It sheltered the Stalwarts, Half-Breeds, and Mugwumps of the Spoils system. The Rockfellers and the Goldwaters. The Tea Partiers, Neo-conservatives, and Paleo-conservatives. But who are today’s occupants? I think there are four: traditional conservatives, the tech-right, populists, and post-liberals. And on AI, they each want something different. 

Today, beneath the facade of MAGA, the conservative-libertarian remnants of the pre-Trump fusionist consensus: veterans of the “Culture War,” Obamacare, and Iraq, though thinned by defections from its more liberal-wing (Bill Kristol, Mitt Romney, Liz Cheney, etc), is still a powerful faction. The ‘Reagan Republicans’ exist on a  spectrum from business-friendly institutionalists like John Thune to Religious Right leaders like Mike Johnson, but they all share a view of American politics frozen in the 1980s, where rhetoric about ‘Evil Empires’ and ‘Common sense governance’ has not yet grown stale. On AI policy then, these traditional Republicans are the swing faction because they are ideologically committed to pursue some goals that the other factions see as mutually exclusive: they want to beat China, protect children online, respect markets, and also respect state sovereignty against an overreaching federal government. Traditional Republicans lack any one framework for how they might accomplish these goals; instead, they’re opportunistic—when an issue arises that might help strengthen one of these priorities, they’re happy to notch a win for their side, but often with little consideration for how their dealings might impact their other interests. Despite their lack of media saviness, old age, and ossified opinions, traditional conservatives are the sleeping giant of the Republican Party; the majority of Republicans in both legislative chambers fit under this label, as well as a sizable chunk of agency personnel. And thanks to the work of the Federalist Society, almost all conservative judges in federal courts are aligned with this old-school Republicanism. Traditional Republicans also still make up a relevant portion of think tank staff at the established right-of-center organizations like the Heritage Foundation, American Enterprise Institute, and the Hoover Institute. Indeed, the recent resignations at Heritage prove the ubiquity of this type even after its much-trumpeted transformation into a more populist organization under Kevin Roberts.

The other old-school wing of the Republican Party with some heft—due in large part to recent reinvigoration from Silicon Valley—is the libertarian faction. The libertarian tendencies now packaged as the trendy ‘Tech Right’ are hardly fresh. Indeed, conservative intellectual Russell Kirk complained in 1981 about the Republican party’s tendency to ally with “metaphysically mad” libertarians, declaring that “if one has chirping sectaries for friends, one doesn’t need any enemies.” But the fusionism of Goldwater and Reagan ensured libertarians, from Ron Paul, who at times certainly resembled the “harmless political lunatic” maligned by Kirk, to Paul Ryan’s suit-and-tie deficit scaremongering. Today, when one thinks of a libertarian, it’s simultaneously hard to look past the infamous “driver's license” debate from the 2016 Libertarian Party Convention or figure out exactly why everyone was so concerned about healthcare—the “individual mandate,” “death panels,” etc, and the deficit back in 2013. Both branches of libertarianism represented by the above figures have burned themselves out: go to a Libertarian Party convention today, and you’re more likely to see Donald Trump speaking than a delegate stripping naked on stage. And since Republicans failed to repeal Obamacare, no one has heard from Paul Ryan, and neither side of the aisle wants to talk or even think about the humdrum technicalities of healthcare policy, much less fixing the deficit. No, today, while the think-tanks (Cato, Mercatus) have stayed the same, the rank and file libertarian wing of the Republican Party is drawn from fresh blood: the so-called ‘Tech Right.’ It would take a PHD thesis to answer the question: how did Silicon Valley become libertarian? And is there any relationship between those old-school libertarians and these guys (Yes, but it’s complicated and best summed up by considering their shared love of Bitcoin and hard drugs) Libertarian AI policy, like all of their opinions for better or worse, follows directly from their abstract principles: for a number of different reasons (epistemological, moral, pragmatic, etc) they converge on the belief that everyone should be free to do as they please unless they are directly harming others, and so AI development should be completely unencumbered unless you can show tangible harm. Some of these Silicon Valley libertarians have also strangely taken up an eschatological view of AI, which gives them further grounds to believe that AI should not be regulated, or at least that it should be left to them—the most component actors in their worldview—to do it. For a brief moment, the tech-right appeared—through DOGE—to be remarkably well-positioned in the second Trump administration. That mirage had passed, and today, the remnants of the tech-right are sheltered in obscure but powerful executive branch agencies like the Office of Science and Technology Policy and in think tanks like FAI, the Mercatus Center, and the Cato Institute. 

Today, the tech-right’s most vehement opponents are the populists. Now, the populist wing of the Republican Party is a bit like its shadow—whatever issue voters that otherwise are solidly Republican don’t feel like is represented in the think tanks and glitzy appointments becomes the territory of the “populist.” This, of course, makes it difficult to trace the development of whatever the modern strain of populism is, because “populism” simply depends on whatever issues the masses care about, which Beltway Republicans are willing to overlook at any given time. For example, the “populists” of the 2010s were the fiscal hawk Tea Partiers until the brief ascendance of Sarah Palin and Paul Ryan, and bear very little resemblance to the immigration plus tariff “populism” of today. Perhaps then, with Trump’s two victories, it is not quite right to talk about “populists” since once again the concerns of the conservative base have been integrated back into the party, or perhaps it would be better to say transformed the party. However, the moniker is still useful insofar as, since the Republican party is a coalition, there are individuals who purport to represent this great mass of voters—many of whom were ex-Democrats—who brought Trump his victories. The self-labelling of populists is something like “MAGA Republicans.” The (up until recently) Majorie Taylor Greenes, Boeberts, and Paxtons of the party. On AI policy, like gene-editing and cloning that were perfect AM radio fodder in the 2000s, the populists unsurprisingly take a completely negative view: the elites and megacorporations threaten the sanctity of the American worker and want to throw us into a “New World Order” where we eat bugs and get paid a UBI to do nothing. This does not seem like a particularly appealing future, so the populists are deeply suspicious of the AI industry—especially those closest to home on the tech-right. The populists have a few outspoken representatives in the House and some weight in the think tank world (Heritage under Robert’s, America First Policy Institute), but their real strength is the personal connection that firebrands like Loomer and Brannon have with Trump—so the power of the populist wing ebbs and flows with Trump’s mood. 

The populists’ high-brow cousins are post-liberals. Post-liberalism is a tricky term: in its most technical sense (and in an American context) it refers to a group of largely Catholic intellectuals like Adrian Vermule, Patrick Deenen, and Sohrab Ahmari who viciously attack the individualism and vacuity of modern liberal politics and instead seek its replacement with a more communitarian polis. But any discussion of post-liberalism, at least in DC, will quickly become spiced with comments about JD Vance, Hungary, and Peter Thiel. It suffice to say that ‘post-liberalism’ is a flexible category with room for a set of divergent ideologies—intellectuals who share a distaste for the social impact of free markets (overlapping with populists in how, for many, the demise of the Rust Belt ‘revealed’ liberalism’s flaws) and the way liberalism has hollowed out meaning beyond an individual’s whims. Now, perhaps this is all good and interesting political philosophy, but what does it mean in terms of policy? Post-liberals are united against mainstream conservatives and the tech-right in their broad willingness to use the government to intervene in the market, whether through antitrust, tariffs, or subsidies (to increase the birthrate). So on AI policy, the post-liberals and the populists are unambiguous in their support for laws regulating artificial intelligence. Post-liberals would like significant oversight over AI by the government so it can be used to support the ‘common good.’ (Of course, what this means will depend greatly on who you ask.) The most powerful post-liberal in the government is JD Vance, and so naturally their influence is relegated to the executive branch: the office of the vice president and the State Department, as well as upstart think tanks like American Moment. 

DC insiders have loosely grouped Republicans into these factions for the past few years, but it wasn’t until last May that their generally banal infighting turned into a major conflict—and it was all about AI policy. 

——————————————————————————————————————————

On May 11th, the House Energy & Commerce Committee majority staff circulated a memo ahead of a May 13 markup on budget reconciliation (“One Big Beautiful Bill”). Buried in the Communications subtitle was a provision that said:

“No state or political subdivision may enforce any law or regulation regulating artificial intelligence models, artificial intelligence systems, or automated decision systems during the 10-year period beginning on the date of enactment.”

The next day, Democrats on the committee released a press release decrying the “moratorium” on states regulating AI, and the story was quickly picked up by the Washington Post and The Guardian. A week later, the moratorium provision was included in the reconciliation bill passed by the House, with barely any negative commentary whatsoever from Republicans. A notable exception was an angry tweet from Majore-Taylor Greene, which foreshadowed the furor yet to come. 

At this point, the only object of controversy among Republicans was whether the moratorium could be removed from the reconciliation bill under the Senate’s Byrd Rule, which keeps the Senate from passing non-budgetary related laws in the reconciliation process. To address this, Ted Cruz, Chair of the Senate Commerce Committee, introduced a revised version of the moratorium on June 5th, which would prevent states from receiving federal broadband funding if they imposed state-level AI regulations—satisfying the conditions of the Byrd Rule. 

It wasn’t until June 18th that cracks began to emerge: Tennessee Senator Marsha Blackburn (who had criticized the bill with MTG in early June) appeared alongside the Tennessee Attorney General and attacked the moratorium on behalf of the music industry—emphasizing that the moratorium would void Tennessee’s 2024 Elvis Act passed which prohibited AI voice imitation without a musician’s consent. 

From this point forward, the floodgates opened: opinion pieces in National Review (traditional conservative), Compact (postliberal), and RedState (populist) assailed the moratorium. The split among Republicans was not just dissent from a few marginal (MTG) or localistic (Blackburn’s opposition on behalf of Nashville) figures—now it was a battle playing out in media outlets, press conferences, and the floor of the Senate. Blackburn, MTG, and Josh Hawley—who answered, ‘it better be out’ when asked if the moratorium would be stripped in the Senate bill—represented three of the four Republican factions: traditional conservatives, populists, and post-liberals, now united against the fourth: the tech-right, pushing for the moratorium. 

The tech-right tepidly counter-attacked. On June 27th, the Cruz-chaired Commerce Committee issued a press release compiling almost a dozen statements from various public interest groups and think tanks in favor of the moratorium. Opinion pieces in favor of the moratorium appeared in Newsweek, Barron's, and RealClearPolicy. Influential tech-right luminaries Marc Andresswitz and Palmer Lucky threw their full support behind the provision. 

It would not be enough.

A last-minute compromise between Blackburn and Cruz—where laws related to copyright or children’s online safety were exempted—fell apart less than a day after it was put together when a group of populist activists led by Steve Bannon called up her office in a fury. 

The next day, during the reconciliation session ‘vote-o-rama,’ the Senate passed a motion to strip the moratorium from the reconciliation bill 99-1. Not even Cruz voted for it. 

————————————————————————————————––––————————

The real irony of the moratorium fight was that you knew it was over the moment you heard about it. Once it made the front page, the game was up. The tech-right’s attempt at a regulatory decapitation strike exposed their political naivety the same way their boldly planned and ill-executed DOGE adventure did—once the bad press mounted and ‘Big Balls’ was a household name, it didn’t matter how many well-meaning hedge fund executives were willing to come out of retirement to help out: it was all over.

It’s not enough to say that they underestimated their opponents, though it’s easy to recognize why they brushed off the provocations of Marjorie Taylor Greene and Laura Loomer without realizing that the force of their rhetoric — a remarkable set of cliches and turns-of-phrases organically sharpened since the days of Rush Limbaugh, Project Veritas, and Breitbart — was worth more than every penny Marc Andreessen and Ben Horowitz gave to Trump’s 2024 campaign ($5 million). No, for all their talk about ‘status’ and ‘moving fast and breaking things,’ the tech-right still simply did not understand how power worked in Washington: how one should bury friendly clauses deep into the reconciliation bill without arising suspicion, how to work reporters, and how (though, of course, a smooth political operator should never let it reach this point) to tell the right story about a contentious topic so that there’s something in your narrative for everyone in the GOP (and their constituents) to love. 

The death of the moratorium, like most dramatic political events, was overdetermined: the traditional conservatives didn’t want it (you can easily imagine Mike Johnson breathing a sigh of relief when it was killed, regardless of his regime rhetoric during the fight), the postliberals didn’t want it, and the populists didn’t want it. And each one of these factions had their own set of powers they could use to try and stop it: the populists (Greene, Loomer, Bannon) had a direct line to Trump and they used it to keep him from stepping in and saving the moratorium, the postliberals in the executive branch, both in Vance’s Office and (from the rumors I’ve heard) in the Office of Science and Technology, kept both of these powerful forces sidelined as well, and the traditional conservatives—the majority of Senate Republicans, had, without the interference of the Executive Branch, the votes to send the moratorium to its grave. 

Of course, these factions are not all-powerful. Cruz’s compromise with Blackburn was a facile attempt to peel off traditional conservatives from the anti-AI coalition by addressing the practical concerns they’d raised: copyright and child safety. He was probably wrong to think that this would be enough to mollify their concerns: that these conservatives didn’t have the same reflexive anti-AI stance as the postliberals and populists, hence Blackburn’s quick turn-around back to her original coalition, but if Cruz had succeeded, the moratorium would almost certainly have been passed. The postliberals and populists alone simply didn’t have enough support in Congress for them to do anything about it—they needed the traditional conservatives. The power of the populists ebbs and flows with how Trump is feeling on any particular day—their direct line to Trump is their power, but given his fickleness, it’s also a liability, as we’ve since seen with the Epstein Files and MTG’s resignation. The postliberals fill the ranks of political appointees, especially those on the younger side, and have an ally in the vice-president, but they lack any congressional support whatsoever beyond Josh Hawley ever since Marco Rubio became Secretary of State. 

And vice versa: if, for whatever reason, the populists or postliberals had supported the moratorium and brought in Trump or executive agencies, the traditional conservatives likely would have folded rather than risk a primary challenge for going against the Trump administration policy. 

Indeed, the moratorium battle made the traditional conservatives’ weakness especially apparent: they depend on a consensus about technology among Republicans that no longer exists. Gone are the Bush administration years when the Republican Party was largely in lock-step about regulating new technologies— when traditional conservatives could always expect to extract concessions from the libertarian wing of their party without a fight, an era best exemplified by the Children’s Online Privacy Act and the Genetic Information Nondiscrimination Act, as well as the (often populist-created and trad-con channeled) opposition to human genetic engineering, cloning, and other technologies with unclear cost-benefits (and often significant tail risks) which began to emerge around the turn of century. Democrats were writing articles the day after the AI moratorium rule was proposed, but it wasn't until more than a month later that traditional Republicans had a spokesperson (Marsha Blackburn) and angle (music copyright). 

——————————————————————————————————————————

But what can a careful look at the moratorium fight and the strengths and weaknesses of these factions tell us about future Republican AI policy?

For the next four months after the AI moratorium’s defeat, the AI war was a frozen conflict. But in December, the fighting restarted: after Republicans declined once again to pass an AI moratorium, this time added onto the annual defense authorization bill, President Trump issued an executive order directing the Justice Department to challenge state-level AI regulations as unlawful. And days earlier, Trump had announced he was lifting export controls on NVIDIA's H200 chips so they could be exported to China. A coalition of traditional conservatives, populists, and postliberals won the moratorium battle in July, but now the tech-right has counterattacked. What does the future hold?

Three political institutions have the power to determine American AI policy: the executive branch, Congress, and state legislative assemblies. The tech-right has no sway over state-level Republicans and even if it did, regulations tend to operate at the level of the lowest common-denominator—the most restrictive municipality sets the tone for everywhere else (see ‘The State of California says this causes cancer’ labels or the endless negotiating that tech companies are forced to do with the EU and highly interventionist countries like Russia, China, and Turkey)—so they would be at the mercy of Democratic states like California and New York. So, neutering state-level power over AI is the most important move for the Tech-right. If they cannot do this, they will never get the policies they want. But the new strategy: an executive order, is either just a scare-tactic or a Hail Mary — though it includes all the relevant carve-outs (child safety and possible discretion for deep-fake and copyright laws) needed to make it palatable to traditional conservatives, the legal issue it turns on (the dormant commerce clause) will likely require overturning the recent Supreme Court precedent set in National Pork Producers Council v. Ross (2023).

Even though the executive order effort is legally dubious and will likely fail if used too aggressively, if they act aggressively and/or Republicans hold onto Congress after the upcoming midterms, the tech-right will probably get some form of federal pre-emption of AI laws by adding a rider to a budget or reconciliation bill with carveouts that traditional Republicans can tolerate and using the executive branch as a proverbial stick to pressure states from imposing onerous AI laws. But they will certainly not get the same broad carve-out they tried to slip by in June. 

The broader future of Republican AI policy will depend on whether the anti-AI coalition can hold itself together in the face of pressure on traditional Republicans to peel off for the sake of a robust China policy and on populists from the ever-fickle Trump. The postliberal faction alone is worthless—its value comes from being able to launder the anti-elitist conspiracy theorizing of the populists to a more intelligent audience and convince traditional Republicans that their social conservatism is irreconcilable with a blanket free-market approach—it is the coalition’s glue, but it has little power alone since, of all the factions, it has the least representation in the government and policy organizations. 

Still, though the tech-right clocked victories in the recent controversy about export controls and, of course, the executive branch’s stance towards state AI bills, the Republican Party’s anti-AI coalition showed a remarkable amount of organic vigor during the moratorium fight, and so it would be a mistake to completely write it off. If the deep inveterate factionalism exposed by the moratorium fight can tell us anything, it’s that power over AI policy in the Republican Party is still very much up for grabs.

Owen is a third-year student at the College.