Instagram cracks down on AI profiles, warns of reach cut for unlabeled accounts

Instagram is making it harder for AI-generated people to pass off as real humans on the platform.

The social media platform on August 31 introduced a new “AI-generated profile” label for accounts that feature an AI-generated person and said it will limit the reach of such profiles if creators do not disclose that they are AI-generated.

Instagram is renaming its existing “AI creator” label to “AI-generated profile” to make the distinction clearer for users. The company said it has heard from people who do not like discovering only later that a profile that appeared to belong to a real person was actually entirely AI-generated.

Under the new system, creators can add the label by editing their profile and switching on the AI-generated profile option. The label will appear on the profile and alongside its content where the individual post or media does not already carry an AI label.

Also Read: Meta removes ads for fraud apps posing as porn after India sounds alarm

The important catch is that Instagram says creators who do not label their AI-generated profiles could lose recommendation reach.

The platform said accounts it detects as featuring an AI-generated person without the label will be notified. They can then check their Account Status to see whether their profile has been made ineligible for recommendations.

Creators have two ways to get their recommendation eligibility back — add the AI-generated profile label or appeal Instagram’s decision if they believe the restriction was made in error.

Instagram is also making a distinction between an AI-generated person and creators who simply use AI tools.

“If you are using AI as part of your creative process, you do not need to add the profile-level label,” the company said. The new rule is specifically aimed at profiles where the person featured is AI-generated.

Instagram has been pushing AI labels for a while

The move is part of Meta’s larger effort to make AI-generated content easier to identify across Facebook, Instagram and Threads.

Also Read: OpenAI's ad business hits $1 billion annualised revenue run rate

Meta started expanding its AI labelling system in May 2024 after its Oversight Board recommended that AI-altered content that could mislead users should be labelled rather than simply removed.

Between October 1 and October 29, 2024, Instagram recorded more than 330 million pieces of content carrying the AI Info label that received at least one user label view. That included nearly 140 million posts, more than 220 million reels and over 24 million stories.

According to Meta, Instagram users generated more than 1 trillion AI Info label views during the same period, with over 60 million label actions, including users clicking on the label to learn more.