How to Opt Out of AI Training on Your Data

Almost every major platform now uses your content to train AI models, and almost all of them opt you in by default. There is no single universal opt-out — you have to visit each platform separately.

Set expectations first: opting out stops future training. It does not remove data already absorbed into a trained model — extracting it (“machine unlearning”) is an unsolved technical problem. Do it anyway; every cycle you sit out is data that stays yours.

The platform-by-platform list

1. ChatGPT — Settings → Data controls → turn off “Improve the model for everyone.” Full guide
2. Google Gemini — Settings → Gemini Apps Activity → turn off, then delete existing activity. Full guide
3. Meta (Facebook, Instagram) — Settings → Privacy Centre → submit the AI objection form. Full guide
4. LinkedIn — Settings → Data privacy → “Data for generative AI improvement” → off.
5. X — Settings → Privacy and safety → Grok → turn off training. Full guide
6. Anthropic (Claude) — consumer chats are not used for training by default; confirm under Settings → Privacy.
7. Adobe — Account → Privacy and personal data → turn off content analysis.
8. Slack — workspace-level setting; ask your admin to opt the workspace out.

If you own a website

Block the AI crawlers in your robots.txt:

User-agent: GPTBot
User-agent: ClaudeBot
User-agent: Google-Extended
User-agent: CCBot
User-agent: PerplexityBot
Disallow: /

Reputable crawlers honour this. It is voluntary, not enforced — but it also establishes that you did not consent, which matters legally.

Deletion, not just opt-out

Opt-out toggles are a courtesy. A deletion request under GDPR or CCPA is a legal obligation. If you live somewhere with those rights, file the request in writing as well — see our state privacy rights guide.