Methodology
How it works
The AI Ideology Test
The test asks you to agree or disagree with 28 statements about the future of AI, plus two bonus questions about AI figures and companies that don't affect your result. Each statement is linked to one or more of 13 dimensions, and pushes your score on each dimension up or down depending on your answer and how strongly the statement relates to it.
| Dimension | Measures |
|---|---|
| Acceleration | How fast should advanced AI be built? (Slow down → Full speed) |
| Regulation | How much control should governments have? (Hands off → Strict oversight) |
| Safety concern | How seriously should catastrophic AI risk be treated? (Overblown → Existential) |
| AI optimism | Will increasingly powerful AI benefit humanity? (Wary → Hopeful) |
| Automation | How comfortable are you with AI replacing human labor? (Protect work → Automate it) |
| AI autonomy | How much independence should AI systems have? (Tools only → Free agents) |
| Human control | Should humans always keep final authority? (Delegate → Humans decide) |
| Open source | Should powerful models be openly available? (Locked down → Open weights) |
| Institutional control | Should frontier AI be run by a few trusted institutions? (Decentralize → Centralize) |
| Post-humanism | Could intelligence move beyond humans? (Human-first → Beyond human) |
| Economic transformation | Should AI rebuild how work and income function? (Preserve → Rebuild) |
| AI rights | Could artificial minds deserve rights? (Just machines → Moral patients) |
| AI impact | How big — and how soon — will AI change the world? (Overhyped → Civilizational) |
Your score on each dimension runs from 0 to 100, where 50 means balanced or unsure. Every ideology has an ideal profile across the same dimensions, weighted toward the ones that define it. Your result is the ideology whose profile is closest to yours. The match percentage reflects that distance: 90%+ is a near-perfect fit, 70–89% a clear one. You also see your runner-up and your ideological opposite.
Scoring happens on our servers. The browser only ever receives the statements, never the weights or the ideology profiles, so results can't be reverse-engineered or tampered with. We tested the scoring against tens of thousands of simulated respondents to make sure every ideology is reachable and no single result dominates.
The AI Revolution Index
- One person, one current result. You can retake the test any time; only your latest result counts toward the live distribution. Older results stay in your private history.
- Verified vs. all participants. Anyone can take the test anonymously. Participants who sign in with Google are counted as verified.
- Minimum group sizes. No percentage is published for any group smaller than 30 people. This keeps small samples from producing misleading numbers, and makes sure no one's individual answers can be inferred.
Votes and rankings
- Voting requires a Google account. Each account has one current vote per question, which you can change at any time.
- Rankings only include people and companies with at least 30 votes. “Most controversial” rewards large, evenly matched camps on both sides.
- Every vote change is recorded in an append-only log, and daily snapshots are rebuilt from it, so trends like “down 5 points this month” reflect real changes of opinion.
Keeping results honest
Public voting attracts brigading, so every vote passes server-side validation, per-account and per-network rate limits, and invisible bot detection. We monitor for unusual voting spikes and clusters of accounts from the same network. Fraudulent votes are invalidated and drop out of every total and historical snapshot. Totals are always computed on the server; nothing a browser sends is trusted.
Limitations
AI Revolution measures the views of the people who take part, mostly viewers of the AI Revolution YouTube channel and people they share it with. That is a self-selected, AI-interested audience, not a representative sample of the general public. Treat the results as what this community believes, not what everyone believes.
Estimated ideologies for public figures and companies are our editorial judgment, based on public statements, actions and policy positions, and are labeled as estimates wherever they appear.
Questions or corrections? See About.