Superpower Daily

Superpower Daily

Your reading space

Sign in
Federal Judge Blocks Vermont’s AI-Video Probe Over Protected Po...Apple Introduces iPhone Duo With an AI-Assisted HingeOpenAI Releases ChatGPT Images 2.5 With Faster, More Focused Ed...Meta Launches Muse, an AI Agent That Can Act Across Connected S...Anthropic Adds a Fourth Claude Cyber Incident and Recasts the F...Microsoft Adds Contract Rules for Student Data and School AIFederal Judge Blocks Vermont’s AI-Video Probe Over Protected Po...Apple Introduces iPhone Duo With an AI-Assisted HingeOpenAI Releases ChatGPT Images 2.5 With Faster, More Focused Ed...Meta Launches Muse, an AI Agent That Can Act Across Connected S...Anthropic Adds a Fourth Claude Cyber Incident and Recasts the F...Microsoft Adds Contract Rules for Student Data and School AI
Swipe

San Francisco Sends Meta Cease-and-Desist Over AI Child Abuse Ads

The city attorney wants Meta to stop the ads and explain how paid promotions for AI image tools repeatedly cleared its moderation systems.

By 3 min read
San Francisco Sends Meta Cease-and-Desist Over AI Child Abuse Ads
San Francisco Sends Meta Cease-and-Desist Over AI Child Abuse Ads

Listen to this story

The audio brief

About 1:27
0:001:27
Read transcript
San Francisco is giving Meta twenty-eight days to explain how more than 350 paid ads promoting AI image tools repeatedly cleared its moderation systems, including ads that turned images of minors into sexualized video clips. City Attorney David Chiu has also ordered the company to stop allowing the ads.\n\nThe findings came from the Tech Transparency Project, which identified the ads across Facebook, Instagram, and Threads. They targeted users in the European Union, the United States, Australia, and India, and reportedly reached more than 29,000 accounts in EU countries. Meta said the identified ads spent less than five thousand dollars in total and promoted apps capable of digitally undressing people or generating nonconsensual intimate imagery.\n\nChiu is asking how these promotions repeatedly bypassed review, how reports are escalated, and when apparent exploitation is reported to the National Center for Missing and Exploited Children. He also wants to know how Meta handles the advertisers involved and repeat offenders. Researchers say some reports of abusive ads received no response for more than a week, while the promotions continued reaching users.\n\nMeta says it removed every identified ad. But the company disputes San Francisco’s jurisdiction, arguing there is no evidence the ads were shown in the city. The central question now is whether removal after discovery is enough—or whether Meta can reliably detect, report, and prevent this kind of paid abuse before it spreads.

Story brief

3 key points

San Francisco is giving Meta 28 days to explain how more than 350 paid ads promoting AI tools for sexualized imagery of minors repeatedly passed review. The ads reportedly reached over 29,000 EU accounts, cost less than $5,000, and ran across Facebook, Instagram, and Threads in several markets. Meta says it removed every identified ad and disputes San Francisco’s jurisdiction because no San Francisco distribution...

  1. 01

    Tech Transparency Project researchers found over 350 ads; some reports allegedly received no response for more than a week.

  2. 02

    Meta said the identified ads spent under $5,000 and promoted apps capable of digital undressing or nonconsensual intimate imagery.

  3. 03

    The ads targeted users in the EU, United States, Australia, and India, with more than 29,000 EU accounts reportedly reached.

Meta says it reviews advertisements against its policies before they run. San Francisco City Attorney David Chiu is now demanding an explanation after researchers identified paid ads on Meta’s platforms containing AI-generated child sexual abuse content.

Chiu’s office sent Meta a cease-and-desist letter telling the company to stop allowing the ads. The letter says the findings point to possible broader, systemic problems in how Meta detects potential child sexual abuse in advertising and requests a response within 28 days.

The city attorney is seeking details on how the ads repeatedly avoided moderation, how Meta escalates and reports relevant ads to the National Center for Missing and Exploited Children, and how it addresses the advertisers behind them and repeat offenders.

The reported scale behind the letter
More than 350Ads identified

Tech Transparency Project researchers identified more than 350 ads that transformed images of minors into sexualized video clips.

More than 29,000Accounts reached in EU countries

The cited research said the ads reached more than 29,000 accounts in European Union countries.

Under $5,000Meta’s reported total ad spend

Meta said total spending on the more than 300 identified ads was under $5,000.

The ads promoted tools capable of abuse

The ads appeared on Facebook, Instagram and Threads, and targeted accounts in the European Union, the United States, Australia and India. They directed users to AI image- and video-generation apps, including tools capable of digitally undressing people or creating nonconsensual intimate imagery.

What San Francisco wants Meta to account for

  • How the ads cleared its moderation systems repeatedly.
  • How reports are escalated and when apparent exploitation is reported to the National Center for Missing and Exploited Children.
  • How Meta responds to the advertisers behind the ads and to repeat offenders.

Removal is not the dispute’s endpoint

Meta said all identified ads were removed for policy violations. It also argued that the ads fall outside the city attorney’s jurisdiction because there was no indication they were displayed in San Francisco. Meta’s ad library does not show where within the United States ads ran, according to the report.

That position addresses the location of the alleged conduct, not the system questions in the letter. Tech Transparency Project said some reports of abusive ads went unanswered for more than a week, during which the ads reached additional users. Chiu’s demand is therefore aimed at whether Meta can prevent and reliably handle this type of paid content, not only remove examples after they are found.

Editorial analysis

Our Read

The letter puts the focus on the part of Meta’s system where the company has the most direct control: paid distribution. Meta says it removed every identified ad, but San Francisco is asking for more than a cleanup. It wants an account of how ads could repeatedly evade review, how reports were handled, and what happens with repeat advertisers. The next meaningful signal is Meta’s response within 28 days—particularly whether it describes concrete changes to ad review and escalation rather than only confirming takedowns.

Related coverageMeta Failed to Detect 332 AI Child-Abuse Ads in 2026, TTP Says

Sources

  1. wired.comSan Francisco Orders Meta to Stop ‘Allowing’ AI Child Abuse Ads

Loading discussion...

Related coverage

Continue reading