Nothing in American politics is more powerful — or more arbitrary — than the moment a complex policy question stops being a question and becomes a tribal badge. Climate, Israel, vaccination, free trade: in each case, a position that was once held comfortably across party lines hardened into a litmus test — not because the underlying facts changed, but because political elites signaled, and voters sorted.
Climate is the textbook case. It was not that Americans uniformly believed in global warming thirty years ago and then divided. Nobody thought much about it at all. Early evidence trickled in during the 1990s with neither party owning a position. The sorting gained momentum only when Al Gore made it his personal cause, and once one party's most prominent figure claimed an issue, the opposing party's incentive structure did the rest. Today, the partisan gap on climate is the widest on any major policy issue — 66 percentage points separating Democrats who call it a critical threat from the 16 percent of Republicans who agree. Note how contingent this was: Republicans, as the party of rural America — farmers, ranchers, people who live on the land — had every natural reason to own environmental stewardship. Energy independence through renewables is a national security argument tailor-made for hawkish conservatism. But once "the environment" got filed under "government regulation," the sorting was irreversible.
Gore tried the same move with the internet and failed — and the reason he failed tells us something important about which issues can be sorted and which cannot. The internet was an equal-opportunity technology that delivered obvious, immediate benefits to everyone who touched it. By the time Gore claimed credit, millions of people were already using email and buying books on Amazon. You cannot capture a technology that your opponents are already using every day. Paul Krugman famously dismissed the internet's economic significance in 1998; but his own readers were reading him online, which made the dismissal self-refuting in real time. The internet spread too fast and too broadly for any party to own it.
Climate change, by contrast, was abstract. It offered no immediate personal benefit. It required trusting models and scientists rather than one's own daily experience. That abstraction left a vacuum — and in American politics, vacuums get filled by partisan entrepreneurs. Gore filled it from the left; the right responded by defining itself in opposition; and within a decade, a scientific question had become a cultural identity.
The question for AI, then, is: does it follow the internet pattern (too useful for anyone to own) or the climate pattern (abstract enough to be captured)?
We will state our prediction plainly: AI will follow the climate pattern, and its boosters will be overwhelmingly Republican. The reason is that AI's benefits, while real and enormous, are not distributed the way the internet's were. The internet gave everyone email on day one. AI's productivity gains are accruing first to a specific type of person — entrepreneurial, self-directed, comfortable with technology — and threatening another specific type: credentialed professionals whose value came from gatekeeping expertise. That asymmetry is partisan fuel.
The alignment is already forming. The Trump administration has embraced AI enthusiastically, repealing Biden-era executive orders and positioning deregulation as competitiveness against China. A growing number of Democrats are putting AI regulation at the center of their 2026 campaigns, tapping into parental anxieties about children's safety and worker displacement. Over $100 million from AI-linked political groups is flowing into the 2026 midterms. The structural pattern is unmistakable: Republicans defaulting to entrepreneurial freedom, Democrats to institutional oversight.
Consider how the Israel parallel reinforces the point. A decade ago, no Democrat would have imagined that vocal support for the Jewish state would become uncomfortable in progressive circles. The percentage of Democrats who view Israel positively has fallen from 34 percent in 2024 to 13 percent in 2026. AIPAC, once a bipartisan powerhouse, is now toxic among Democratic primary voters. The sorting took about ten years from first cracks to near-total polarization. AI is on a similar timeline.
Or recall how the Obama campaign in 2008 treated social media as a progressive technology — a tool for organizing young voters against entrenched power. Democrats were proud Facebook users right up until 2016, when they decided social media was responsible for Trump. The technology hadn't changed. The perceived beneficiary had. We predict AI will follow the same arc: a technology that progressives initially claimed as their own, until they notice it empowering the wrong people.
The coalition logic makes the sorting self-reinforcing. Tech libertarians — Musk, Andreessen, Thiel — join naturally with small-government conservatives. On the other side, the anti-business left unites with the constituencies most directly threatened: teachers' unions (who see ChatGPT as an assault on pedagogy), university professors (whose monopoly on credentialed expertise is dissolving), and Hollywood (at war with AI over intellectual property and synthetic performances). These are not abstract coalitions; they are already visible in campaign spending and candidate platforms.
And here is the deeper economic irony. AI's productivity gains will flow disproportionately to exactly the pro-entrepreneurship, small-business types who form the Republican base. A sole proprietor who can now do the work of a ten-person marketing department; a small manufacturer who automates logistics without hiring consultants — for these people, AI is an unambiguous gift. Meanwhile, trade workers — plumbers, electricians, HVAC technicians — will thrive in a world that replaces cognitive labor but cannot send a robot to fix your furnace. The plumber who voted Republican and now charges premium rates while former middle managers compete for dwindling white-collar jobs may become the most pro-AI worker in America, even if he never touches a chatbot.
The most interesting constituency to watch may be religious Americans. Christians who believe in the soul have a theological resource that secular progressives lack: a principled reason to be unafraid. The Thomistic tradition draws a hard line between genuine intelligence — which requires a soul, and therefore life — and mere calculation, however sophisticated. If human dignity is grounded in something machines categorically cannot possess, then even the most powerful AI is just a tool. The Vatican's recent guidelines define AI as "a gift of human creativity, which itself is a gift from God." That is not the language of existential threat.
But religious conservatives will have their own AI concerns — and those concerns point in a very different regulatory direction from the progressive one. Evangelicals and Catholics share a deep suspicion of centralized power over information and discourse. The well-documented left-leaning bias of every major LLM gives this suspicion real teeth. Religious voters will push not for slowing AI down but for keeping it open: free speech protections, anti-bias requirements, resistance to corporate monopoly. Their regulatory impulse will be decentralize and de-bias, not pause and restrict. That aligns far more naturally with the Republican embrace of AI than with the Democratic skepticism. Do not expect Christians to join the "slow this technology down before Skynet kills us all" caucus — that is an overwhelmingly secular anxiety rooted in a materialist worldview that cannot explain why humans are special. Christians already know why.
For technology professionals who have spent careers assuming their politics and profession point in the same direction, the next few years will require uncomfortable rethinking. The Democratic Party is becoming the party that regulates, restricts, and fears the thing you build for a living. The Republican Party is becoming the party that celebrates it — for reasons having nothing to do with your own. Just like social media after 2016, the technology will not have changed. The perceived beneficiary will have.
Welcome to the sorting machine.