AI Policy

Too Many Zeros uses artificial intelligence — this is how we keep it honest.

The Short Version

Every profile on this site is written by AI and reviewed by a human before it is published. We use AI to do the heavy lifting of research and drafting. A human checks it and decides what goes live. All our AI generated pages state that up top.

Also see our Terms of Use.

How a Profile is Made

1. Research. An AI assistant (Claude, made by Anthropic) searches public sources for a person's documented actions — both good and bad — across philanthropy, labor, environment, politics, innovation, corporate integrity and personal integrity.
2. Sourcing. Every factual claim is linked to a public source you can click and verify yourself. If we can't source it, we don't say it.
3. Writing. The AI drafts the profile — summary, evidence, and analysis — in plain language.
4. Rating. The AI assigns a rating (😇 Good, 😈 Bad) to each piece of evidence. The rating is opinion and commentary — not a statement of fact, and not financial, legal, or investment advice.
5. Grading. The AI categorizes the evidence and calculates a grade (A, B, C, D, F). The grade is a summary of the discovered evidence — not a statement of fact, and not financial, legal, or investment advice.
6. Human review. A person reviews each draft before publication.

Our Goals

• Only documented claims. We write about what the evidence shows. We do not speculate, and we do not make claims about the absence of evidence ("nothing was found") — the internet is vast and we cannot prove a negative.
• Everything is cited. If a statement isn't backed by a linked source, it shouldn't be there. Tell us if you find one.
• These are living drafts. Profiles are early and will improve over time as new information emerges. Wealth figures, in particular, change constantly.

What AI Gets Wrong - And How We Improve It

AI can misread a source, over-generalize, or state something with false confidence. We reduce this risk by requiring a clickable source for every claim and by having a human review each page. We will not always get it right. When we're wrong, we want to fix it.

When corrections are needed, a human guides AI to make those corrections. They set up new rules and verification steps to prevent and catch mistakes. This way, every error is a chance to make future content better.

Conflict of Interest

This project's AI is Claude, made by Anthropic. When we profile people affiliated with Anthropic (for example, its founders), that is a conflict of interest: an AI assessing its own maker cannot be treated as neutral. Those profiles carry a visible conflict-of-interest disclosure and receive additional human review before publication.

Corrections & Feedback

Found an error, a broken source, or something unfair? We want to hear it. Feedback.

This policy is a draft and will evolve.