Meta hit on multiple fronts over AI harms: nudify-app CSAM ads, invasive Meta AI prompts, and biometric training lawsuit

San Francisco's city attorney ordered Meta to stop running ads with AI-generated child sexual abuse content after researchers found 350+ such ads, while Michigan and Florida AGs and Australia's eSafety regulator scrutinize Meta; separately, Meta is fixing…

A Tech Transparency Project investigation found that Meta served more than 350 ads containing AI-generated child sexual abuse content promoting AI nudify apps, some built from images of real minors including a European royal family member. One CSAM ad ran from a Facebook page posing as the Church of Jesus Christ of Latter-day Saints, which Meta disabled only after TTP flagged it. Reports differ on the ads' reach: Ars Technica says they reached over 29,000 people in the EU and 6,800 in the UK with about 80% of ads shown in the US, while WIRED says they reached over 29,000 accounts across the EU, US, Australia and India. San Francisco City Attorney David Chiu issued a cease-and-desist on September 9 demanding Meta stop 'allowing' the ads and answer within 28 days about moderation failures; Meta says all ads have been removed, most had fewer than 200 impressions with total spend under $5,000, and argues the ads ran outside the city attorney's jurisdiction. Michigan and Florida attorneys general are investigating, Senator Mark Warner demanded action, and Australia's eSafety regulator requested information from Meta at a senior level. Australia also released draft Digital Duty of Care legislation that would require platforms to offer users over 16 a non-algorithmic feed ('My Feed, My Way'), extend under-18 protections to social media, games, apps and AI chatbots, give the eSafety Commissioner removal-notice powers over nudify apps, and impose penalties up to AU$109.2 million (US$78.6 million); the bill is expected in Parliament before Christmas. Separately, Meta said it 'missed the mark' and fixed Meta AI after Instagram user Kalie Robins showed it suggesting 'Who is the child passenger?' and piecing together her daughters' ages and location from her and relatives' posts, including surfacing a photo she says was deleted years ago; Meta AI is embedded in Facebook, Instagram, WhatsApp and Messenger. A proposed class action filed in federal court in Chicago alleges Meta extracted biometric data from Facebook and Instagram photos without consent to build the unreleased NameTag face-recognition system for smart glasses — code found in the Meta glasses AI companion app downloaded more than 50 million times before its removal — and to train the Emu and Muse Image generative models. The suit invokes Illinois' Biometric Information Privacy Act, seeking $5,000 per intentional or $1,000 per negligent violation for a class potentially numbering in the millions dating to…

• Tech Transparency Project found 350+ Meta ads containing AI-generated child sexual abuse content promoting AI nudify apps, some using images of real minors including a European royal family member
• A fake Facebook page posing as the Church of Jesus Christ of Latter-day Saints ran a CSAM ad that Meta disabled only after TTP flagged it
• Ad reach figures conflict across reports: Ars Technica reports 29,000+ people in the EU and 6,800 in the UK with ~80% of ads shown in the US; WIRED reports 29,000+ accounts across the EU, US, Australia and India
• San Francisco City Attorney David Chiu issued a cease-and-desist requiring Meta to respond within 28 days; Meta argues the ads ran outside the city attorney's jurisdiction
• Meta says all ads have been removed, flagged ads averaged fewer than 200 impressions, and total ad spend was under $5,000
• Michigan and Florida attorneys general are investigating; Senator Mark Warner sent Meta a letter demanding an end to the ads; Australia's eSafety regulator requested information from Meta at a senior level
• Australia's draft Digital Duty of Care would offer users over 16 a non-algorithmic feed option, protect under-18s across social media, games, apps and AI chatbots, give the eSafety Commissioner nudify-app removal-notice powers, and set…
• Meta AI suggested 'Who is the child passenger?' beneath a user's video and compiled her children's ages and location from her and relatives' posts, including a photo the user says was deleted years ago; Meta says it fixed the issue