Instagram will show fewer people AI profiles that do not label themselves

Instagram said on August 31 that accounts built around AI-made people must carry a label saying “AI-generated profile”. Accounts that do not will be shown to fewer people. The new label replaces the “AI creator” label that arrived in May. Instagram says the new wording makes it clearer that the person on screen does not exist.

The rule covers profiles whose main subject is a made-up person. Virtual influencers fall into that group. It does not cover creators who use AI to edit photos, tidy up captions or make graphics while still appearing as themselves. Creators turn the label on in their profile settings. It then shows on the profile and next to posts, unless the picture or video already says it was made with AI.

Instagram enforces the rule through recommendations. If it spots an AI profile without a label, it tells the account through Account Status and stops suggesting the account to others. In practice that cuts the account off from people who do not already follow it, in Reels and in Explore. Creators can appeal, and if the appeal succeeds the account is suggested again. Instagram says people do not like finding out later that a profile they took for human was made by a machine.

TechCrunch found a dating app running unlabelled AI influencers

Virtual influencers were a small corner of the market for years. Cheap AI tools have turned them into a spam and fraud problem. TechCrunch points to a dating app that ran unlabelled AI male “influencers” on Instagram earlier this year. It also points to health accounts fronted by AI-made doctors selling supplements. For brands, any campaign built around a made-up person now needs the label from the first day. Without it, the reach the brand paid for disappears.

The change also comes as new law arrives. The AI Act, the EU’s new rules for artificial intelligence, says people must be told when they are dealing with AI-made content. Those duties started to apply on August 2. Platforms are under pressure to show they can check this on a large scale. Meta is under extra scrutiny after a settlement worth billions of dollars with US states this summer over child safety.

Instagram has not said how accurate its AI detection is

Instagram has not said how accurate its detection is, or how quickly reach comes back once a label is added. Agencies that run virtual talent should check their lists now. Adding a label is free. Losing reach in the middle of a paid campaign is not. The bigger open question is whether the label will later apply to single AI-made images posted by human creators. Those stay outside the rule today.