Does Instagram use your photos to train AI? What Meta's own pages say

The Morse teamPublished 8 min read

As of November 1, 2026, Instagram does use your photos to train Meta's AI if they're public. Meta's privacy policy lists "public content you or others create or share, like posts, comments or audio" among what it uses to develop its AI, and its Privacy Center names "posts or photos and their captions" as a source of training data (Meta privacy policy, Meta Privacy Center). Photos on a private account, your direct messages, and photos you never post sit outside that, with exceptions covered below.

We read Meta's privacy policy (effective July 23, 2026), two Privacy Center pages on generative AI, and six Meta and Instagram newsroom posts on November 1, 2026. Every fact links to the Meta page it came from. Meta deserves credit for two things: people in the EU and UK get a form to object, and its Privacy Center admits plainly that you can end up in training data without having an account at all.

Instagram AI training at a glance

Read on November 1, 2026.

What you shareUsed to develop Meta's AI?ControlSource
Photos, captions, and comments on a public accountYesMake the account private, or object (EU and UK only)Meta
Posts on a private accountNot listed: the policy names public contentKeep the account privateMeta
Direct messages with friendsNo, unless someone in the chat shares them with Meta AIDon't bring Meta AI into the chatMeta
Your chats with Meta AI on InstagramYesNone in the USMeta
A public post you later deleteNot used for future training, but the model isn't changedDelete the postMeta
You, in a friend's public photoCan be, even if you have no accountA request form for third-party informationMeta
Other people making AI images from your public postsA separate setting, on by defaultSharing and ReuseInstagram

What exactly does Meta take from Instagram?

Public posts, comments, captions, and audio. The privacy policy's section "Developing and improving AI at Meta" says the purpose is to "develop and improve AI at Meta for Meta Products and for third parties," and lists two kinds of content from you: public content, and your interactions with Meta's AI features (Meta privacy policy). The Privacy Center adds information from the internet and from third parties that provide content (Meta).

Meta says it doesn't attach what it collects to your account. "When we do get personal information as part of the training data, we don't specifically link this data to any Meta account," the Privacy Center says. It keeps training data "for as long as we need it on a case-by-case basis."

Training on public posts isn't new. In its April 2025 post on Europe, Meta wrote that training on public content "is how we have been training our generative AI models for other regions since launch" (Meta, April 14, 2025).

Does Instagram use photos from private accounts?

The policy doesn't list them. The AI section names "public content," and posts on a private account are seen only by followers you approve. Meta's EU and UK posts describe the same scope: "public posts and comments, or public photos and captions" from adults (Meta, September 13, 2024).

Two things still reach Meta from a private account:

  • What you send to Meta AI. Chats with Meta AI are listed for AI development in the same policy section. We cover those chats in is Meta AI private.
  • Messages someone shares with Meta AI. Meta doesn't train on "the content of your private messages with friends and family" unless "you or someone in the chat chooses to share those messages with our AIs" (Meta). Meta updated its European post on March 27, 2026 to say the same: "we may train our AIs on messages that people choose to share with our AI features" (Meta).

Can you be in Meta's training data without posting anything?

Yes. Meta's Privacy Center says: "Even if you don't use our Products or have an account, we may still process information about you to develop and improve AI at Meta," for example "if you appear anywhere in an image shared on our Products by someone who does use them" (Meta). A private account covers your own posts. Photos your friends post of you follow the audience of their account.

The same page says "anyone can submit requests related to their personal information from third parties being used to develop and improve AI at Meta," whether or not they use Meta's apps.

Setting your account to private keeps your own grid out of public content. A friend's public post with you in it is still public content, so ask friends before they post photos of you from a public account.

Does deleting old posts remove them from Meta's AI?

Only going forward. Meta's page on how its models work says that if you delete "a public post," the model "won't change in that moment," but your post is deleted, "and once it's deleted, it won't be used for future training" (Meta). The same page says Meta "occasionally" retrains models on a new set of information, so a deleted post stays in models trained before you deleted it.

Who can opt out of Instagram AI training?

Adults in the EU and the UK, through an objection form. Anywhere else, the Meta pages we read on November 1, 2026 describe no form for your own public posts.

Where you liveWhat Meta trains onCan you object?Under-18sSource
European UnionAdults' public posts and comments, and interactions with Meta AIYes, through a form linked in Meta's notices; Meta says it honors past and new objectionsPublic data from EU accounts under 18 isn't usedMeta, April 14, 2025
United KingdomAdults' public posts, comments, photos, and captionsYes, "at any time," through a form linked in in-app noticesAccounts of people under 18 excludedMeta, September 13, 2024
United States and elsewherePublic content, under the privacy policyNo form described for your own postsNo exclusion described in the policyMeta privacy policy

For people under 16 in the US, the closest thing to a rule is Instagram's Teen Accounts. Meta said on September 17, 2024 that teen accounts are private by default, that this applies "to all teens under 16," and that those teens need "a parent's permission to change any of the built-in protections to be less strict" (Meta). A teen's posts stay out of public content for as long as the account stays private.

Our guide to turning off Meta AI covers what you can switch off and clear in each Meta app. To compare Meta's rules with other companies', see which AI companies train on your data.

Is the Sharing and Reuse setting an AI training opt-out?

No. It controls whether other people can make things from your posts. Instagram's blog says that "if you don't want anyone to create with your Instagram content using AI features at Meta," you can turn it off in Settings, then Sharing and Reuse, and that the setting is "available on mobile only for now" (Instagram). Instagram's post says nothing about training.

The setting drew attention in July 2026. Meta announced on July 7, 2026 that people could generate images in Meta AI by @-mentioning public Instagram accounts, then removed the feature on July 10: "We've heard the feedback that this feature missed the mark, so it's no longer available" (Meta). Some coverage that week framed the toggles as a training opt-out, under headlines like "Meta uses your Instagram photos to train AI: Opt out now" (SunStar, July 10, 2026). Turning the toggles off still makes sense if you don't want strangers remixing your posts. It doesn't take your public posts out of the privacy policy's AI section.

Does Meta use photos in your camera roll?

Not on Instagram, as far as Meta's pages say. The camera roll feature Meta announced on October 17, 2025 is on Facebook: it suggests edits and collages from photos you haven't posted, and it's opt-in. Meta says "we don't use media from your camera roll to improve AI at Meta, unless you choose to edit this media with our AI tools, or share" (Meta). The privacy policy's AI section doesn't list camera roll content, though the policy lists it for other purposes, such as safety (Meta privacy policy).

For the same question about Google's photo app, see does Google Photos use your photos to train AI.

How Morse keeps the words you don't post private

Every rule above applies because what you share sits on Meta's servers, under the same account that holds your chats with Meta AI. Nothing you type in Morse is posted anywhere, and Morse doesn't train AI on your content, sell your data, or show ads, so there's no objection form to find and no setting to flip.

Use it for the parts of Instagram you'd rather keep to yourself: drafting a caption before it goes public, planning a post, or asking a question you wouldn't want tied to your profile. Your chat history and Studio images are stored in your browser, encrypted, not on Morse's servers. You pick a privacy mode for each chat:

  • Private, for drafts you'd rather no company kept. It's where a new chat starts. Open-weight models like DeepSeek, Kimi, and GLM, plus xAI's Grok, run here, and Morse has zero data retention agreements with its inference, search, and other providers, so they keep nothing. Where a provider confirms zero retention on each reply, Morse checks it and refuses a reply that doesn't.
  • E2EE, for anything about other people that no one should read. Your browser checks the secure enclave's hardware attestation before anything leaves your device, then encrypts the message so only that enclave can read it, and our servers relay text they can't read. The enclaves are run by third-party providers, not by Morse, and the guarantee rests on the enclave hardware and attestation working as designed. E2EE chats come with Pro, $19 a month.
  • Anonymous, to ask GPT, Claude, or Gemini without your Instagram identity attached. Our proxy removes your account, email, and IP address and sends the message under Morse's own account. The provider sees the message without knowing it's you and may keep it for a limited time to check for abuse, under its own policy, so leave out names.

History locks when you sign out, and unlocking takes two keys: one that never leaves your device and one our server gives you at sign-in. Clearing your browser deletes your chats, and Morse can't recover them, so keep anything you need.

Send a message from the Morse home page without an account, or start free with no card needed.

Common questions

Does posting "Goodbye Meta AI" stop Instagram from using my photos?

No. None of the Meta pages we read on November 1, 2026 mention a post or story that changes how your content is used. The controls Meta describes are your account's audience, the Sharing and Reuse settings, and, in the EU and UK, the objection form.

Can Meta train AI on my Instagram Stories?

Meta's privacy policy, effective July 23, 2026, lists "public content you or others create or share" for developing AI and doesn't single out Stories. A Story from a public account counts as public content, and one on a private account doesn't.

If I object in the EU, does Meta delete what it already trained on?

Meta's April 2025 post promises to honor objections but doesn't say it will retrain models. Its Privacy Center says a model "won't change" when you delete something it learned from, so an objection works on what comes next.

Can I find out whether Meta trained on my photos?

Not from the pages we read on November 1, 2026. Meta's Privacy Center lists where training data comes from but describes no tool that shows which posts went into a model, and Meta says it doesn't link the personal information in training data to any account.

Ask what you actually want to ask.

Over 30 models, characters, and images, with every chat encrypted and kept on your device. Free to start.

Start a private chat

Read next