More people saw these AI-generated ED "remedies" than read The New York Times

🤥 Faked Up #41: AI videos flog bogus erectile dysfunction cures to millions on TikTok; deepfake detectors fail latest test; and Meta ran 660 more ads for AI nudifiers.

Faked Up paid subscribers get access to my annotated collection of studies on digital deception. As of this week, the library includes 48 entries:

Faked Up academic library

This is a regularly updated collection of academic studies cited in Faked Up or that have otherwise caught my eye. It currently includes short descriptions of 44 academic studies and systematic reports organized in five clusters: 1. Prevalence and characteristics of misinformation 2. Effects of fact-checking interventions 3. Prevalence, effects,

Faked UpAlexios Mantzarlis

Today's newsletter is a ~5 minute read and contains 49 links.

NEWS

Italian newspaper Il Foglio launched an AI-generated edition; its first fact check left me skeptical. Draft legislation transposing the EU AI Act in Spain would place hefty fines on companies that fail to label AI-generated content. A Chinese directive also requires labeling of AI content. US academic centers studying misinformation are under renewed legal attack. Instagram is testing AI-generated comments, because why the hell not. Jim Jordan called on Sundar Pichai to pull a Zuckerberg and forsake moderating misinformation on YouTube. One in ten British secondary school teachers claim their students have created malicious deepfakes. Antivaxxers are trying to interpret RFK Jr.'s flip flop on the MMR shot. Meta is trumpeting in Australia the same fact-checking program that it trashed in America. AI-generated YouTube videos of fake showdowns between EU politicians reach millions. A new preprint found that Community Notes that link to fact checks are deemed more helpful than ones that link to anything else studied.