Meta just proved how fast a privacy backlash can move: it launched an AI feature that let anyone turn your public Instagram photos into AI-generated images without asking you, and killed it two days later once people noticed. That’s the entire lifespan of the @-mention tool inside Muse Image, Meta’s new AI image generator, launched July 8, dead by July 10.
Muse Image is the first product out of Meta Superintelligence Labs, running on the Muse Spark model that replaced Llama earlier this year. It’s built into the Meta AI app, Instagram, and WhatsApp, and lets people generate images from text prompts. The part that got Meta in trouble was simple: inside Meta AI, you could @-mention any public Instagram account and use that account’s photos, reels, and feed videos as source material for an AI image, no permission required.
If your Instagram account was public, you were opted into this by default. Someone could tag your profile, generate an event invitation or a mockup or whatever else using your face and your posts, and you’d never know it happened. Meta also confirmed that turning off the setting after the fact wouldn’t delete any AI content already generated from your photos; the damage, such as it was, stuck around.
The only way to opt out required digging into a settings menu most people never open: Sharing and Reuse, then toggling off “Posts” and “Reels” under “Allow people to create with and reuse your content on Instagram and with AI features at Meta.” Reaction on forums and social media was sharply negative. Users called the default-on setting “extremely disappointing,” arguing it should have been turned off by default and that there should be better privacy controls. Some users reported switching their accounts to private entirely in response, with one self-described photographer calling the situation atrocious.
Facing this backlash, Meta removed the @-mention feature altogether just two days after launch. According to a statement shared by Puck’s Dylan Byers, the company said:
Earlier this week, we announced that one way for people to generate images in Meta AI is by @-mentioning public Instagram accounts that they want to reference. Our intent was to provide a useful creative tool and to give people control over whether their public content could be referenced in this way. We’ve heard the feedback that this feature missed the mark, so it’s no longer available.
Statement from Meta:
Earlier this week, we announced that one way for people to generate images in Meta AI is by @-mentioning public Instagram accounts that they want to reference. Our intent was to provide a useful creative tool and to give people control over whether their…
— Dylan Byers (@DylanByers) July 10, 2026
This is not the first time that Meta has pulled such a move with complete disregard for user consent before using their data. Meta admitted in 2024 to scraping all public Facebook and Instagram posts made since 2007 to train its generative AI model, a revelation made during a public inquiry in Australia that applies globally. A year earlier, a separate Meta AI privacy incident saw the company fail to indicate to users what their privacy settings were as they posted. The pattern of shipping AI features with default-on settings that expose user content has become a recurring theme in Meta’s product rollouts.
