ChatGPT has wrong information about me: what to do

If ChatGPT has wrong information about you—wrong job title, wrong company, mixed up with someone else, or made-up details—you’re not alone. AI models sometimes "hallucinate" or pull from outdated or conflicting sources. The first step is to see exactly what each major model says; then you can decide how to fix or correct it.

This article covers why ChatGPT gets it wrong, how to document the errors, and what actually helps: measuring first, then improving sources and disambiguation where possible.

Why ChatGPT has wrong information about you

ChatGPT (and Claude, Gemini) use training data and sometimes live retrieval. Wrong information can come from: outdated or incorrect web sources, name collision (another person or company with a similar name), or the model inventing details. Until you see the answers and, if needed, a diagnostic of where they come from, you’re guessing.

A Scan shows you what OpenAI, Anthropic, and Google each say about you in one report. If the wrong information appears in one model but not others, the cause may be model-specific. If it’s consistent, the cause may be shared sources or entity resolution—a Snapshot can clarify.

What to do when ChatGPT says the wrong thing about you

Don’t assume you can "edit" ChatGPT directly. You can’t. What you can do is: (1) Know what it says—measure with a Scan. (2) Understand why—get a Snapshot that shows retrieval, sources, and entity resolution. (3) Improve the inputs—where improvement is plausible, a Blueprint defines a canonical description, disambiguation, and where to publish what. You execute; we don’t control model outputs.

For personal data removal (e.g. under GDPR), OpenAI offers a "Right to be Forgotten" request via their Privacy Portal. That’s separate from fixing wrong facts; it’s for requesting removal of personal data from ChatGPT responses. For wrong or misleading facts, the path is measure → diagnose → improve sources.

Next step

Get a Scan to see what ChatGPT, Claude, and Gemini currently say about you. If the information is wrong, a Snapshot will explain why and whether a fix is realistic. From there you can correct sources, add disambiguation, or request removal where applicable.