On 31 August Instagram renamed its label for AI profiles and tightened the rules. Accounts run by an AI-generated persona without a label will no longer be recommended in Reels and Explore. In this way the platform, for the first time, regulates identity rather than content.
Renaming the label and tightening the rules
The old label was called “AI creator”. Instagram introduced it in May 2026 as an optional tag and aimed it at accounts that regularly share AI content. The new “AI-generated profile” label targets something else. It marks profiles where the presented person is itself AI-generated or substantially created by AI. It appears in the feed, Reels, Stories, search results and comments. Meta has been building on this labelling policy since 2024.
Anyone who adds the label themselves loses nothing. The reach of such accounts stays unchanged. An account that the system labels by mistake can appeal through Account Status on Instagram.
Where the line for mandatory labelling lies
Instagram set out the reason precisely in its announcement. “As generative AI becomes a bigger part of how people create, we've heard that people don't like seeing a profile that seems human, only to find out later that the person featured is AI-generated.”
So a single question decides it. Is there a presented person, or was the entire person generated by a model? Creators who use AI only as a tool for retouching photos, writing captions or making graphics fall outside the scope. They do not need the label, and the penalty does not apply to them.
The new rules target the second group. AI-generated personas, virtual influencers and synthetic brand ambassadors operate as business models of their own, with their own logic of earning. Many of them have run on Instagram without any clear labelling.
The EU AI Act and the impact on the European market
Instagram's announcement came a month after Article 50 of the EU AI Act took effect. From 2 August 2026 it requires providers of generative AI to mark synthetic content in a machine-readable way. Systems that were already on the market before that date have a transition period until 2 December 2026.
The platform and the regulation aim at the same thing from different sides. Instagram protects users from disorientation. The EU AI Act addresses the entire supply chain of AI systems. How exactly the two approaches will intersect is not yet clear. In particular, the EU sanctions framework (up to €15 million or 3% of global annual turnover for breaching transparency obligations) will apply to platforms depending on how national regulators interpret it, and that interpretation has not yet settled.
Comparison with the approaches of other platforms
Instagram is not the first to set conditions for AI profiles. YouTube has required labelling of realistically altered or synthetic content since 2024. Since May 2026 it has also detected AI content by itself, without manual reporting, and a violation risks a warning or demonetisation.
TikTok has required labelling of AI-generated faces, voice clones and deepfakes since 2023. Since January 2025 it has automatically read C2PA metadata. Under the same standard, Meta began applying automatic “Made with AI” labels on Instagram and Facebook as early as the start of 2024.
But the subject of regulation differs. YouTube and TikTok target a specific video or clip, while this time Instagram targets the profile as a whole. The enforcement differs too. The platform does not delete the post, it only limits its reach, and for AI profiles it reduces the visibility of the whole account to people who do not follow it.
Data show why platforms are reaching for identity itself. A study by the Pew Research Center found traces of AI authorship on 35% of websites created after ChatGPT launched. On social networks the share of synthetic content is higher still. Technical detection is not enough at this point for a user to know who they are actually talking to. Labelling identity, not just content, is the answer to exactly this.
Comments