I’ve spent much of the last 3 years helping thousands of the progressive advocacy and political community navigate the AI landscape.
Earlier this summer, I created a field guide for myself to better understand the AI space and navigate it more clearly. Developing it has given me a deeper understanding of how AI-aligned power and money are shaping our society and future.
With the growing dialogue around AI this week, it’s critical to have an explainer that helps people understand the language and players, so here it is. I hope it helps.
The guide starts with definitions of the six current factions (as I’ve defined them, borrowing from previous great guides like Michel Justen's Guide to the AI Tribes, with input and reviews from dozens of people in and outside of the AI space. The deck has a full set of sources.
The guide to the six AI factions are here. Future posts will dive into other parts of the guide.
Note about groupings: I know organizations and groups aren't monoliths, and there are divisions within orgs and these groups, and some will not like how they’re grouped. Please give us feedback on where we’re getting it wrong or right- or leave a comment.
Figures are as of Sept. 2026.
1. Accelerationists / Tech Right
What they believe. AI delivers abundance. Regulation slows progress, entrenches incumbents, and hands the future to China. Worth noting this is not the same thing as the Abundance movement, despite the shared vocabulary.
Who and how much. Marc Andreessen and a16z, David Sacks as White House AI czar, the OpenAI orbit, Y Combinator. This is the best-funded faction in AI politics by a wide margin. Leading the Future has pulled in $125M or more from a16z and OpenAI’s Greg Brockman, and spent roughly $8M against Alex Bores (NY assemblymember, RAISE Act author) in the NY-12 primary, which he ultimately lost. Innovation Council Action has pledged around $100M. They are allied with the administration through the AI Action Plan and the December 2025 preemption executive order, and their goal is singular: one federal framework that preempts state AI law.
How to spot them. Listen for “little tech,” which frames startups and open source developers as the victims of regulation, and “regulatory capture,” which is the claim that safety rules are an incumbent plot to lock out competitors. “Doomer” is used as an insult.
Splits/Surprises: Do not treat this faction as monolithic with the labs. Meaningful daylight exists between OpenAI’s leadership and its staff, and between individual executives.
2. AI Safety
What they believe. AGI is arriving soon and could be catastrophic. Regulate frontier models before it is too late.
Who and how much. Eliezer Yudkowsky, LessWrong, and MIRI on one end. The effective altruism network and Coefficient Giving (Formerly Open Philanthropy), which has moved $530M or more into AI risk work. Future of Life Institute, Center for AI Safety, PauseAI, Blue Dot. Anthropic-adjacent, in ways that are contested inside the faction itself. Politically, this camp is transpartisan and increasingly well funded: Public First Action, which is Anthropic-linked, spent roughly $11M backing RAISE Act author Alex Bores in the NY-12 race. They back state safety legislation, which puts them at odds with state preemption.
How to spot them. “p(doom)” is a common term, meaning your estimated probability that AI causes human extinction. So is “alignment,” the technical problem of making systems pursue their designers’ intended goals. Both signal a research-first worldview where safety is an engineering problem rather than a political one.
The internal fight worth knowing. Yudkowsky and MIRI purists want to shut it all down, while the Coefficient Giving and Anthropic wing wants to build carefully - and believes that Anthropic is the best available hope.
Organizing Efforts: The Safety movement has started to embrace (read: fund) tactics and approaches from social movements, launching organizations to use community organizing and mobilization efforts to engage the broader public in the safety movement. So far we have Irreplaceable (center left), Humans in Control (trans-partisan), Humans First (center-right), and Evitable. Interesting to track the different ways each group frames its argument and its theory of change.
3. AI Ethics and Bias
What they believe. Present-day harms come first. Algorithmic bias, exploitation of data workers, copyright, water and energy consumption. Existential risk talk actively distracts from harms that are already landing on real people. This is the TESCREAL critique.
Who and how much. Timnit Gebru at DAIR, Meredith Whittaker at Signal, Joy Buolamwini at the Algorithmic Justice League, the FAccT conference, and unions in the middle of acute harm fights. They are aligned with digital rights groups and parts of labor, and they shaped Biden’s AI Bill of Rights.
This faction has enormous intellectual influence and currently no significant political money, at least that's visible in any public disclosure.
Words and Beliefs. “Stochastic parrots,” from the 2021 paper, meaning large language models mimic language without understanding it. And “TESCREAL,” the acronym bundling transhumanism, Extropianism, singularitarianism, cosmism, Rationalism, effective altruism, and longtermism into a single critique of the safety camp as quasi-religious. Both moves deflate AGI hype and pull the conversation back to the present.
Splits/Surprises: They want product liability for harms to kids, and they oppose age verification mandates. If you assume child safety is a simple left-right issue, this faction might scramble some assumptions.
4. AI Populists
What they believe. AI as a technology may be good or bad. What matters is who controls it. The current trajectory concentrates wealth and power, and accelerates oligarchy and autocracy.
Every other faction here is defined by its answer to “what is AI?” This one is defined by “who has power?” That makes it a political-economy faction rather than an AI-prediction faction, which is why it doesn't fit neatly on other maps. A lot of the progressive political world may find themselves here.
The term was popularized by writer Jasmine Sun, who frames it somewhat differently than I do.
Who and how much. Guardrails Alliance, Roosevelt Institute, Groundwork Collaborative, Tech Oversight Project, Demand Progress, Americans for Financial Reform, and a growing set of progressive organizations. AI Now increasingly operates here. Humans in Control and Irreplaceable sort of sit in both this camp and safety camp. The Guardrails Alliance PAC launched in June 2026 with roughly $5M raised toward a $15M goal, funded by tech workers, AFT, AAUP, and parent groups. Its first buy was $250K backing Bores. Tomorrow Together launched in early September, 2026 with a reported $10m to spend on midterm House candidates who support child-safety and data-privacy regulations and limiting data center construction.
How to spot them. A focus on the Tech Oligarchy and billionaires controlling infrastructure, media, and now government. And “guardrails,” meaning baseline enforceable rules with no AGI prediction attached. These folks are largely pro-regulation but don't take a position on whether AGI is coming.
Their strategy. Near-term harms like surveillance, deepfakes, and job loss are legitimate concerns and organizing gateways to structural demands about power (see: safety-aligned organizing efforts outlined above). The data center moratorium is the (current) main people-power lever, and the token tax is the insider version of the same play. This coalition includes many diverse interests, so don’t assume a unified position.
This is the accountability wing of the progressive movement arriving in AI politics, and is increasingly a political home for progressives concerned about wealth and power concentration.
5. MAGA Populists
What they believe. Big Tech cannot be trusted. The concerns are job loss, child safety, data centers, surveillance, and AI as a threat to family and religious values.
Who and how much. Steve Bannon, Marjorie Taylor Greene, Senator Josh Hawley. Their power is coalitional and it is inside the Republican Party. They helped kill the state preemption moratorium 99 to 1 in the Senate.
How to spot them. Child safety is their sharpest AI issue, and they are comfortable with age verification, which puts them opposite the AI Ethics camp on remedies while agreeing on the problem. They are increasingly against warrantless tracking. And they will attack their own administration over data centers and utility bills.
Why progressives should care. This is the faction most likely to be an unlikely bedfellow on issues like preemption, data centers, and now on surveillance; the fight is Tech Right versus MAGA base, and it is open war inside the GOP. This is where the traditional left vs right framing of politics shifts.
6. Pragmatists
What they believe. AI is transformative and also normal, like electricity. Both the utopian and the doomer framings are overblown.
Why they matter. They have no identifiable organized political money, but pragmatists get regularly cited in committee hearings and quoted in the technical sections of bills. That makes them the faction most likely to shape what a law actually says, and how regulatory frameworks are developed.
Who and how much. Arvind Narayanan and Sayash Kapoor of AI Snake Oil at Princeton, and a lot of academics. No organized movement or PAC/political spending as of now.
How to spot them. “AI as normal technology.” Case by case on regulation, skeptical of the China race framing.
N.B.
I know that not everyone will agree with how their group is framed or lumped with others. I look forward to feedback on what we’re getting wrong - and when things are moving so fast, I’m sure its a number of things. Comment below, send a message here or on our website
Future posts: Outlining where each group falls on different issues like china, open source, child safety, etc.
Let me know what other topics you hope we dive into
So grateful for you mapping this out. We have needed a shared lexicon as we try to wrangle/navigate/comprehend the brave new world we seem intent on creating for ourselves.
This was very insightful. Thanks for writing this.