Instagram’s new AI can turn your photos into deepfakes – here’s how to opt out
What happened
On July 9, 2026, Meta launched an artificial intelligence image generator on Instagram. The tool can take photos from public accounts and use them to create new, AI-generated images — effectively turning your real posts into raw material for deepfakes. Privacy experts quickly raised concerns, and coverage from outlets including The Guardian, Yahoo Finance UK, and inkl has drawn attention to the lack of upfront consent.
Under the default settings, most public Instagram accounts are automatically opted in. That means any photo you’ve shared publicly can be fed into the AI model. Private accounts are not affected in the same way, but the company has not given clear details on whether private content is wholly excluded from training.
Why it matters
The risk goes beyond the abstract unease of having your photos “used by AI.” Deepfake imagery can be weaponised for harassment, scams, blackmail, or reputational damage. For children and teens who often have public accounts without realising the implications, the danger is especially acute. UK schools have already been advised to remove pupils’ online photos as the AI blackmail threat grows.
Even if you trust Meta not to misuse your images today, the data used to train generative models can persist indefinitely. Opting out now prevents your photos from becoming part of a training set that could be applied in ways not yet imagined.
What you can do
The opt-out is buried in Instagram’s settings, but it only takes a few taps once you know where to look. Follow these steps:
- Open the Instagram app and go to your profile.
- Tap the menu (three lines in the top right), then Settings and privacy.
- Scroll down to Privacy and select AI Preferences (the exact label may vary slightly by region).
- Toggle off Allow use of your photos for AI training.
That’s it. The setting should save immediately. Note that this only covers future use of your images — photos that may already have been used cannot be recalled, because model training is not reversible.
Additional steps to protect yourself
- Review the privacy of past posts. If your account is public, consider switching to private, or at least archive older photos you don’t want in the public pool.
- Limit who can share or repost your content. Go to Privacy > Posts and turn off Allow reshares to stories if you don’t want your images spreading.
- For parents: check your child’s account settings. Make sure their account is set to private and that the AI opt-out is enabled.
- Be cautious about posting high-resolution, identifiable images of yourself or family members, especially if you keep your account public.
None of these steps will stop someone from screenshotting your public photos and using them with third-party AI tools, but they remove Meta’s direct access to your content for their own generator.
A note on limitations
The opt-out is not a silver bullet. Companies like Meta often update terms and settings, so check back periodically. Also, private accounts may still have their images used in aggregated, anonymised training — Meta has not confirmed otherwise. The most reliable way to keep your photos out of any AI system is to post as little as possible and to keep your account private.
Sources
This article draws on reporting from The Guardian (“Instagram’s AI image generator alarms privacy experts”), Yahoo Finance UK (“Instagram has just allowed anyone to alter your images with AI. Here’s how to opt out”), and inkl (“Meta’s New AI Can Turn Instagram Photos Into Deepfakes”). For the school warning, see The Guardian, “UK schools should remove pupils’ online photos as AI blackmail threat grows, say experts” (May 2026).