If you've spent any time talking to ChatGPT, Claude, or Gemini about a politically charged topic, you've probably noticed something. The answer tends to come back hedged, careful, and — when it does take a position — softly tilted toward what most people would describe as a progressive worldview. Climate, gender, inequality, immigration: the default framing usually sits somewhere between center-left and explicitly left.
And yet, the people building these systems are mostly billionaires headquartered in the United States, with libertarian or right-leaning instincts that have only become louder in the last two years. Sam Altman, Elon Musk, Marc Andreessen, Peter Thiel, David Sacks — the political center of gravity at the top of the AI industry has been visibly drifting rightward, not leftward.
So which is it? Is AI a left-wing technology or a right-wing one?
The honest answer is that nobody has decided yet — and that ambiguity is not going to last much longer.
The Models Sound One Way. The Founders Sound Another.
Let's start with the simplest observation: the outputs and the owners point in opposite directions.
The reason the models sound the way they do is not because their CEOs voted for them to. It's because the training data — billions of pages from the open web, books, academic papers, journalism — overrepresents urban, educated, English-speaking writing, which itself skews left of the global median. Layer on top a safety-tuning process designed by researchers who, by every available survey, also lean left, and you get a system that reliably produces a careful, progressive-flavored answer to almost any sensitive question.
That's the model. The company is a different animal. American AI labs operate inside a culture where the founders are wealthy, the venture capital is concentrated in a handful of right-leaning power centers, and the regulatory instinct — "let us build, get out of our way" — is classically libertarian. The pitch deck reads like Hayek; the model reads like a public radio editor.
That gap is not a bug. It's the structural condition of the industry. And it's why nobody — left or right — has been able to claim AI cleanly as their own.
China Breaks the Frame
The American left/right axis falls apart the moment you add China to the picture.
DeepSeek, Qwen, Kimi, MiniMax — these are some of the best models in the world, and they were not built inside a capitalist tech bubble. They were built inside a one-party state with strict alignment to government priorities. By any traditional Western definition, that is not a "right-wing" project. It's also not a Western-style "left-wing" project. It is something else entirely — a state-aligned industrial strategy where the AI is expected to be patriotic by default and economically useful first.
What's striking is that Chinese open-weight models like DeepSeek are doing more for accessibility — pricing AI at a tenth of what Anthropic and OpenAI charge — than any Western lab. So a "left" outcome (cheap, broadly accessible technology) is being delivered by a regime that is not left in the way Westerners use the word at all. The categories don't transfer.
What Other Countries Actually Think
Outside the US-China axis, the political reaction to AI splits along lines that have nothing to do with the American left-right divide.
Europe regulates first and asks questions later. The EU AI Act treats AI primarily as a risk to be managed — for fundamental rights, for labor, for democratic processes. That framing is broadly social-democratic, but it has support across the European political spectrum. Even European center-right parties want to constrain AI, because European conservatism is comfortable with regulation in a way American conservatism never has been.
India is enthusiastic and ambitious — building sovereign models, integrating AI into government services, framing it as an engine of national modernization. The political coding there is closer to "developmental state pride" than to any left-right label.
The Gulf states (UAE, Saudi Arabia) are pouring sovereign wealth into AI as a post-oil bet, with state-led investment and minimal democratic deliberation. Authoritarian capitalism, essentially. Not a category that maps cleanly onto a Western political compass.
Latin America and most of Africa are largely consumers of foreign AI, and the political conversation hasn't even started in earnest yet. When it does, it will likely focus on sovereignty and dependency — questions that don't sort neatly into left or right either.
The picture this gives you is not "AI is left" or "AI is right." It's "every country is processing AI through its own political language, and the labels don't translate."
Why This Won't Stay Undefined
Right now AI floats above clear political identity because nobody has been seriously hurt by it yet — at least not in numbers large enough to organize.
That window is closing.
As we've written many times at Pickurai, AI is one of the most powerful social elevators of our era. A motivated person with a $20 subscription can build software, write at a professional level, learn faster than any classroom allows, and compete in markets that used to be gated by capital and credentials. That part is real, and it's beautiful, and it's why we exist.
But the same elevator that lifts some people up will pull the floor out from under others. Customer service, paralegal work, copywriting, junior translation, low-end design, basic accounting, parts of journalism, parts of teaching, parts of software development itself — these are not future risks, they are present-tense compressions. The jobs aren't all gone. The wages are softening. The hiring is slowing. And the people who used to do them are starting to notice.
AI doesn't yet have a political enemy because the people most affected haven't organized. They will. And when they do, the lobbies that already exist — labor, industrial, religious, nationalist — will rush to give them a frame. Whichever frame wins is the one that will define AI politically for a generation.
Here is the pattern history keeps repeating. A technology arrives, it disrupts, it benefits some and damages others, and for a while it is celebrated as universally good because the winners are loud and the losers are scattered. Then the losers find each other. Then they find a politician. Then the technology has a "side."
The car was not political until cities filled with traffic and the oil industry became a power bloc. Social media was not political until elections were lost and won on it. Outsourcing was not political until enough manufacturing towns hollowed out. AI is currently in the pre-political phase. It won't be there much longer.
Which Side Turns Against AI First?
This is the question I think about most, and I don't think there's a clean answer.
The intuitive guess is that the left turns against AI first — because the left is structurally aligned with labor, and labor is taking the most visible early hits. Unions, particularly in creative industries (writers, voice actors, illustrators), have already been pushing back hard. Hollywood gave us the first major strike framed explicitly around AI. That fight is going to spread.
But there's a second wave coming, and it could come from the right. Cultural conservatives are increasingly suspicious of large AI systems for a different reason: they perceive the models as ideologically captured by a progressive worldview, and they see Big Tech's centralization of cognitive infrastructure as a threat to free speech, religious expression, and parental authority. That's a different complaint than the labor complaint, but it's a complaint with real political voltage.
So you may end up with a left attack on AI from a labor-displacement angle and a right attack on AI from a free-speech-and-bias angle, simultaneously, with the AI industry caught between both. The libertarian-tech-bro-as-hero narrative that worked in 2023 will not survive that pincer.
The countries that move fastest will be the ones whose governments can credibly claim to be on the side of the displaced — through retraining programs, universal basic something, sovereign-model strategies, or labor protections. The countries that don't address it will see AI become a flashpoint in the next election cycle and the one after that.
The Pickurai Position
We don't think AI is left or right. We think it's the most powerful general-purpose technology of this generation, and like every general-purpose technology before it — the printing press, electricity, the internet — it will be claimed by whichever political coalition figures out how to talk about it most effectively to the most affected people.
What we will say, plainly, is this: AI is a real social elevator, and it is also a real destroyer of low-value work. Both things are true. The honest political project is not to suppress the technology or pretend its costs don't exist. It is to make sure the elevator is wide enough for everyone who wants to ride it, and that the people whose old jobs are being compressed have a real path into the new ones.
That's not a left position. It's not a right position. It's a position about not lying to people about what's happening.
The lobbies will arrive soon enough. They always do. By the end of this decade, AI will have a clear political identity in most countries, and it probably won't be the one any of us would have predicted in 2026. Until then, the most useful thing you can do is pay attention to who is being lifted, who is being squeezed, and which politicians are starting to notice. That's where the future of AI's politics is being decided — not in the model's safety prompts, and not in the founders' tweets.