How to Opt Out of Llama Model Training (Meta)
AI Privacy Opt-Out Guides — Opt out of AI services' use of your personal information for future training data.
Meta trains its Llama family of models on public posts, comments, photos, and captions from Facebook, Instagram, and Threads accounts — plus content about you posted by other people. This 2026 guide covers the Llama-specific opt-out paths: the right-to-object form for users in the EU and UK, and the third-party-information form that everyone else has to use. It complements our broader Meta data opt-out guide, which covers ad targeting and account-level controls.
Why this matters. Consumer AI products use chats, posts, and account data to train models by default. Once your information is included in a training run, it cannot be removed retroactively. The earlier you opt out, the more of your data stays out. Priwall by mePrism tracks every major AI opt-out so you can act before, not after.
Why opt out of Llama training?
Meta uses your public content across Facebook, Instagram, and Threads to train Llama — the model family that powers Meta AI in chats, on Ray-Ban Meta glasses, and across Meta's ad tools. It also uses posts and photos other people share that reference you, even if you never had a Meta account. Once a Llama training run is complete, the data it consumed cannot be pulled back out. The window to object is before the next training round for your region begins.
Quick facts
How to opt out of Llama training
Check your region. Meta gives EU and UK users a simple objection form. Users in the US and most other regions have to file under a narrower third-party information process.
EU and UK users — open Facebook or Instagram Settings. Go to Settings & Privacy → Privacy Center.
Find the AI section. Click AI at Meta, then choose Right to object.
Fill in the form. Enter your email, your country, and a brief reason (a sentence is fine — "I object to my personal data being used to train Meta's AI models" works).
Submit and save the confirmation. Meta emails a confirmation. Keep it. Objections must be received before a given Llama training round starts for the run to exclude your data.
US and other users — use the third-party information form. Go to facebook.com/help/contact/510058597920541.
Provide evidence. The form only accepts requests to remove third-party personal information from Meta AI outputs — meaning content about you that Meta AI has generated in response to a prompt. Include screenshots of the Meta AI prompt and response that surfaced your information.
Submit separately for each surface. File once for Facebook, once for Instagram, once for Threads if the same information appears in more than one place.
Steps current as of 2026; if the platform has changed its flow, see Meta's Privacy Center and AI at Meta help pages.
Tired of chasing AI privacy settings?
Priwall by mePrism finds your AI opt-out controls across major platforms, helps you turn them off, and monitors for changes — so a single policy update doesn't silently re-expose your data.
Skip the manual work →What to expect after opting out
For EU and UK objections, Meta confirms receipt by email and applies the objection to future training rounds. Data used in prior rounds stays in the models trained on it. For US and other users, Meta reviews third-party information requests case by case and only acts on Meta AI outputs — it does not treat the form as a blanket training opt-out. That's a real gap in Llama's opt-out coverage outside the EU and UK, and one of the reasons ongoing monitoring matters.
Frequently asked questions
Isn't this the same as the Meta opt-out I already did?
Not quite. Our main Meta guide covers ad targeting and account controls. This guide is specifically about Meta using your public Facebook, Instagram, and Threads content to train the Llama models — a separate legal and product process.
Does this cover WhatsApp?
WhatsApp message contents are end-to-end encrypted and are not used to train Llama. WhatsApp account metadata is a separate topic, covered in the main Meta guide.
Does this stop Meta AI from mentioning me in responses?
The third-party information form is the tool for that specifically. If Meta AI surfaces your name, photo, or biographical detail in response to a prompt, use the form with screenshots.
What if I don't have a Meta account?
Meta still trains on public posts that reference you — photos other people upload, comments that name you, tags. The third-party information form is your path in that case.
How often should I re-check?
Meta announces Llama training rounds region by region, sometimes with only weeks of notice. Re-check before any announced round and after any major Meta policy update.
"Meta was the one where I actually didn't know where to start. Priwall walked me through the objection form in the EU, and it caught the third-party form for my US team the same week. Two different flows, one dashboard." — Priwall customer, media
Related AI opt-out guides
Sources checked
MIT Technology Review — How to opt out of Meta AI training
Norton — How to opt out of Meta AI
Meta Help Center — Third-party information about you in Meta AI