TL;DR
- Instagram's stated reason is that people dislike seeing a profile that seems human, only to learn later that the person featured is AI-generated.
- Marketing Oops reported on 1 September 2026 that the account level label is separate from the AI info label Meta applies to individual pieces of content, so per-post disclosure does not cover the profile.
- Social Media Today gave no rollout date, no country list, no figure for how much reach is limited, and no explanation of how Instagram detects an unlabelled AI person.
- In late 2024, reports said Meta planned AI bot personas across its apps that would engage as if human-run; the company does not appear to have moved ahead with it.
- Marketing Oops also reported that a creator who only uses AI as a tool does not need the label, and that applying it correctly does not reduce reach on its own.
Instagram said it will limit the reach of profiles that feature an AI-generated person but do not add its new AI label, according to a report published by Social Media Today on 31 August 2026. The same update renames the "AI creator" account labels that Instagram added in May 2026 to AI generated profile labels, and adds a description noting that the images posted by the account may feature AI-generated characters.
The renaming is the headline, but the distribution sentence is the part with money attached. A label that only sits on a profile is a disclosure setting. A label that decides whether a profile gets normal reach is a ranking input, and that lands on any brand running a virtual presenter, an AI spokesmodel or AI-generated people in its organic creative.
What changed on the label itself
Instagram introduced account level AI labels in May 2026 under the name "AI creator". The 31 August update changes the name to AI generated profile and attaches a description telling viewers that the images posted by the account may feature AI-generated characters. The stated intent is to make it easier for someone scrolling to tell whether a profile is run by a human or is posting AI-generated content.
Instagram explained the change this way: "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. They want to know when a profile features an AI-generated person."
The wording is worth reading carefully because it defines the trigger. The label is about a profile that features an AI-generated person. It is not framed around whether AI touched the editing, the caption or the background. The object of the rule is a synthetic human being presented to an audience.
The sentence that turns a label into a distribution rule
Instagram said it will limit the reach of profiles that feature an AI-generated person but do not add this label. That single line changes the calculation for anyone weighing whether to disclose. Before it, adding the label was a transparency choice with a possible audience cost, since some viewers scroll past anything marked as AI. After it, not adding the label carries a distribution cost that Instagram has now said out loud.
What the report does not contain is any number. There is no percentage, no description of which surfaces are affected, and no statement about whether the limit applies to the feed, Reels, Explore, recommendations or all of them. The mechanic is announced, not quantified, and a marketer treating it as a known penalty size would be inventing the size.
The table below sets out what the source states about the change, with the parts that remain unspecified marked as such rather than filled in.
| Element | Before the 31 August 2026 update | After the update |
|---|---|---|
| Account label name | "AI creator", added in May 2026 | "AI generated profile" |
| Label description | Not described in the report | Notes that images posted by the account may feature AI-generated characters |
| Reach for an unlabelled profile featuring an AI-generated person | Not described in the report | Instagram said it will limit the reach of profiles that feature an AI-generated person but do not add the label |
The profile label and the per-post AI info label are two different things
Thai marketing outlet Marketing Oops, covering the change on 1 September 2026, reported that the AI generated profile label sits on the account and is separate from the existing AI info label that Meta applies to individual pieces of content. One tells a viewer something about who or what runs the account. The other tells a viewer something about a single image or video. The Social Media Today report describes the renaming and the reach limit without setting out that distinction.
That distinction matters operationally because the two can disagree. An account run by a real person who occasionally posts an AI-generated image is a different case from an account whose recurring on-screen presenter does not exist. The report frames the new profile label around the second case, since Instagram's stated concern is a profile that seems human until the viewer learns the person featured is AI-generated. Neither report says how the two labels interact, whether one implies the other, or whether applying content level labels consistently affects the profile level decision.
Marketing Oops, in the same 1 September 2026 coverage, also reported who the label is not for. A real human creator who uses AI as a tool, for retouching images, adjusting captions or generating graphics, does not need the AI generated profile label. The label is for profiles where the person featured is created, or mostly created, by AI.
The same coverage reported that applying the label correctly does not reduce an account's reach simply because it uses an AI persona. The reach limit is aimed at accounts that should disclose and do not. Marketing Oops also reported that an undisclosed AI-generated profile may lose eligibility to be recommended to people who do not already follow it, in Reels and Explore, while existing followers can still see the content.
For brands, the practical read is that per-post AI disclosure on a few pieces of creative is not the same as declaring the account. If your brand account is fronted by a synthetic character, the profile level setting is the one the new rule points at.
The detection question the source leaves open
Instagram has said it will reduce reach for unlabelled profiles that feature an AI-generated person. It has not said, at least not in this report, how it identifies one. There is no mention of a classifier, of provenance metadata such as content credentials, of user reporting, or of any manual review step. There is also nothing about an appeals path for an account that gets the treatment wrongly.
That gap is the part worth watching rather than guessing about. Detection method determines the false positive rate, and false positives on this kind of rule land on heavily retouched photography, stylised illustration and CGI product work that features no synthetic person at all. None of that is addressed in the reported facts. Teams that produce a lot of AI-assisted imagery for content marketing should keep records of how each asset was produced, because the cheapest defence against a wrong classification is being able to show what was made and how.
The backlash this update sits inside
The report places the change against a longer run of events. In late 2024, reports emerged that Meta planned to introduce a range of AI bot personas across its apps that would interact and engage as if they were human-run accounts. The theory behind that plan was engagement: people would get more replies and more likes on their posts, and would feel more seen and heard. Meta did not share further information on the plan, and it does not appear the company moved ahead with it.
The 31 August update runs the other direction. Instead of adding synthetic accounts that behave like people, Instagram is now labelling synthetic accounts more clearly and reducing the reach of ones that do not disclose. The outlet reads this as a response to a growing backlash against AI-generated content and AI bots on social platforms.
Thai marketing outlet Marketing Oops also covered the change on 1 September 2026, so the update reached Thai language trade press within a day. That is coverage, not confirmation of a Thai rollout, and neither the original report nor the local pickup gives a country list.
What the source did not say
The absent details here are the ones a planner actually needs, so they are worth stating plainly.
- No rollout date and no country list. The report gives no timeline for when the renamed labels or the reach limit take effect in any market.
- No definition of how much reach is limited. Social Media Today gives no figure and no description of which placements are affected. Marketing Oops reported on 1 September 2026 that an undisclosed profile may lose eligibility for recommendation to non-followers in Reels and Explore, which says where rather than how much.
- No detail on how Instagram detects an unlabelled AI-generated person, and no mention of an appeals process.
- No statement about whether paid placements are affected. The report discusses profiles and their reach, and says nothing about ads.
- Nothing about Thailand. The policy cannot be described as live in Thailand on the strength of this reporting.
What this means for Thai marketers
Virtual presenters and AI-generated brand characters have been a live creative option in this market for a while, which makes the disclosure question concrete rather than theoretical. The rule as reported keys on a profile that features an AI-generated person, so the accounts most exposed are the ones where a synthetic face is the recurring presence: a virtual influencer account, a brand mascot rendered as a photorealistic human, a spokesperson persona built entirely in generation tools.
Accounts that use AI for backgrounds, product renders, upscaling or retouching, with real people in front of the camera, are not what the stated wording describes. That does not make them safe from a classifier that nobody has described, which is a reason to keep production records rather than a reason to label everything.
The other thing to hold loosely is scope. Nothing in the report says the rule is enforced in Thailand today, and nothing gives a date. A brand deciding whether to declare its account should be doing it on the basis of audience trust and the announced direction of travel, not on a belief that enforcement has already arrived here.
What to check if you run AI presenters in organic content
- Inventory the accounts. List every profile your brand controls where a person appearing in the content is AI-generated rather than filmed. That list is the scope of the rule as it was described.
- Check the account level setting, not just your post captions. Per-post disclosure does not stand in for the profile label, and the profile label is what the reach rule references.
- Write down how each recurring character is produced. If a classifier flags an account wrongly, the useful thing to have is a production record, not an argument.
- Separate organic from paid in your planning. The report says nothing about whether ads are affected, so campaigns running through paid social should not be replanned on an assumption the source does not support.
- Brief the creative team on the distinction. AI used on a background is a different question from an AI person on screen, and the announced rule addresses the second.
FAQ: frequently asked questions
Will Instagram reduce my reach if I use AI-generated images?
The reported rule is narrower than that: Instagram said it will limit the reach of profiles that feature an AI-generated person but do not add the AI generated profile label. The stated trigger is a synthetic person featured by the profile, not AI involvement in an image generally. The Social Media Today report gives no figure for how much reach is reduced. Marketing Oops reported on 1 September 2026 that an undisclosed profile may lose eligibility to be recommended to non-followers in Reels and Explore, and that applying the label correctly does not reduce an account's reach on its own.
What is the difference between the AI generated profile label and the AI info label?
The AI generated profile label applies to an account, while the AI info label is the existing label Meta applies to individual pieces of content, a distinction Marketing Oops set out in its 1 September 2026 coverage. The profile label was previously called "AI creator" and was added in May 2026 before being renamed on 31 August 2026. The Social Media Today report does not explain how the two labels interact.
Does this apply to Instagram ads?
The report does not say. It describes labels on profiles and a reach limit for unlabelled profiles that feature an AI-generated person, and makes no statement about paid placements. Anyone planning around ad treatment is working past what has been reported, which matters for teams whose spend sits mostly in Meta ads.
Is the policy live in Thailand?
The source gives no country detail and no rollout date, so it cannot be described as live in Thailand. Thai marketing outlet Marketing Oops covered the change on 1 September 2026, which shows the news travelled, not that enforcement has started in any particular market.
How does Instagram know a profile features an AI-generated person?
The report does not explain the detection method. There is no mention of a classifier, provenance metadata, user reporting or manual review, and no appeals process is described. That is one of the larger open questions in the announcement as reported.
If a synthetic presenter is part of how your brand shows up on Instagram, the sensible next step is a short audit of which accounts and which assets the rule would touch, and what your production records look like. Relevant Audience works with brands on organic and paid social, and that audit is a small piece of work compared with rebuilding a creative line after a reach drop.
