Skip to content

Product1 publisher3 min readPublished

Britain's ad regulator bans five AI app ads for sexualising women

The ASA ruled against ads for AI companion and video tools running on Meta platforms, including one it found portrayed a character under 18 sexually. None of the advertisers answered its enquiries, and Meta had already pulled the ads.

The Product Desk · Product desk

Illustration accompanying Britain's ad regulator bans five AI app ads for sexualising women

What happened

  • The ASA banned five ads uploaded to Meta platforms for AI companion and video generating apps, saying they had objectified and sexualised women, including through harmful gender stereotypes.
  • Two ads for the AI portrait tool Nexaipic encouraged users to "upload your crush and let AI create what you imagine" with sexually explicit videos.
  • Meta has said it removed more than 344,000 ads for nudifying apps across Facebook and Instagram between November 2025 and January 2026.

Compiled by The Product DeskSomething wrong?How this is made

Why it matters

  • precedent There is now a published UK ruling on the demo asset, so an AI tool can be lawful and still lose its paid social because of what the creative shows happening to a woman's image.
  • constraint Because none of the five advertisers engaged with the regulator, the working penalty is whatever Meta's ad review does plus the published ruling. Against an anonymous developer, that is about as far as the ASA gets.
  • exposure Teams advertising legitimate AI image and video tools sit inside the same monitoring net, since the ASA's system scans live ads on its own initiative instead of waiting for a complaint.
  • decision Growth teams using an "upload your crush" hook have to decide whether the conversion is worth a named ruling that stays online after the ad comes down.

Somebody at each of these apps chose the demo clip. For the app the ASA identified only as KH31 DD22, the clip showed a clothed woman being turned into a video in which she was topless, and the watchdog said that suggested the app could be used to create content exposing women's bodies [5]. The PictoPop ad used sexualised imagery to demonstrate the tool, including portraying the woman in a submissive position [6]. Ads for Rusto AI encouraged users to make sexualised videos from photographs [7].

It is the ad that gets judged. The ASA said the five banned ads had depicted women irresponsibly and were likely to cause serious or widespread offence [8]. The test applies to the creative asset: what the demo shows happening, and whose picture the hook asks the viewer to supply. Nexaipic's two ads asked for the viewer's crush [4].

The legal position moved underneath the investigation. The UK outlawed apps for creating non-consensual deepfake nude imagery a few months ago, after anger over X's chatbot Grok complying with requests to nudify images of women on the platform [9]. The ASA said the law had changed during its investigations, that ads identified in future would be considered in light of both the law and its own rules, and that where ads are illegal it may notify platforms and law enforcement [10].

None of the apps' creators responded to the ASA's enquiries [11]. Meta had already removed the ads for breaching its own rules before the ASA acted [12]. For an advertiser that never answers a letter, the enforceable part of this runs through the ad platform and through the ASA's AI-powered Active Ad Monitoring System, which proactively searches for ads that might break UK rules and found these ones [13]. "We take a firm line on ads that are likely to cause serious or widespread offence and these ads are a clear example of why that matters," said Nicky Baker of the ASA [14].

Five rulings sit against a much larger volume. Meta has said it removed more than 344,000 ads for nudifying apps across Facebook and Instagram between November 2025 and January 2026 [15]. Across the 92 days in that window, that is roughly 3,700 ads a day [1]. Alexios Mantzarlis reported in early September that an investigation found 11,336 ads on Meta's platforms claiming they could turn images of clothed people into sexually explicit videos, most of which Meta removed automatically [16]. Many of them promoted domains or publishers that his site Indicator had previously flagged to the company, he said [17]. Meta declined to comment and pointed to action it has already taken on nudification apps, including new technology to detect ads and legal action [18].

The BBC's account gives the count as five ads, then itemises at least six: one for Animcha, two for Nexaipic, one for KH31 DD22, one for PictoPop and further ads for Rusto AI [2].

If you are the person signing off paid social for an AI image or video tool, two questions carry the risk. Whose image does the ad tell the viewer to upload, and what does the demo show being done to that image. An ad that answers "someone you know" and "clothes come off" fails on the creative whatever the app's own terms permit. There is a third checkpoint, the one Animcha hit: sexualised imagery combined with childlike clothing and objects, on a synthetic character presented as someone users could "customise" and instruct, was enough for the ASA to find the ad portrayed someone under 18 in a sexual way [3].

What to watch

  • Whether the ASA notifies a platform or law enforcement about an AI app ad under the new deepfake law, and which advertiser is first.
  • Whether Meta publishes a nudify-ad removal figure for the period after January 2026 that can be compared with the 344,000.
  • Whether the five named advertisers reply to the ASA or reappear in Meta's ad library under new names.
Loading claim ledger
Loading source directory links
Loading share composer
Loading topic controls
Loading related stories