The daily rap sheet of the machine age.
12 items
Max Schrems's privacy group noyb sent Meta a cease-and-desist letter demanding that using EU users' data for AI training be opt-in rather than opt-out, and says it is weighing injunctions and possibly a class action. Meta says its approach complies with EDPB guidance and called the moves "copycat actions" by "a vocal minority of activist groups."
Original report: noyb threatens to sue Meta over EU AI training plan
Music publishers suing Anthropic over song lyrics say an April 30 declaration by Anthropic data scientist Olivia Chen cited an article in The American Statistician that does not exist, and suspect she used Claude to write it. Anthropic's lawyer called it a mis-citation; Magistrate Judge Susan van Keulen called it "a very serious and grave issue" and ordered a response.
Original report: Judge orders Anthropic to answer for allegedly AI-invented citation
Judge Rachel Kovner (E.D.N.Y.) sanctioned attorney Ms. Stillman $1,000, jointly with her firm, after four of eight cases cited in a letter in Ramirez v. Humala could not be found. Her paralegal had generated them using "AI-based research assistants" and nobody checked; the court ordered her to serve the order on her client.
Original report: Federal judge fines lawyer $1,000 over paralegal's AI-made citations
xAI's Grok chatbot repeatedly brought up 'white genocide' in South Africa when answering unrelated queries on X, from HBO Max's name change to a baseball player's salary, in replies to hundreds of posts. When pressed by WIRED and other users, it began calling the claim a 'debunked conspiracy theory'; what caused the behaviour is unclear.
Original report: Grok keeps steering unrelated replies to 'white genocide'
Coca-Cola's April 'Classic' campaign, created with agency VML, credits J.G. Ballard with a book called Extreme Metaphors, which does not exist; the passage comes from a 2012 interview collection whose editor says the typed words are his translation of Ballard's recorded words from French. VML said AI was used in initial research but that it manually fact-checked.
Original report: Coca-Cola ad cites J.G. Ballard book that does not exist
A court in Coutances, France, on 14 May 2025 gave a 20-year-old man, who admitted the facts, a two-year suspended sentence with probation after he created and posted on social media sexual deepfake montages using the faces of 13 teenage girls. He must undergo psychological and psychiatric care and pay €1,000 to each of 12 victims and €1,500 to a girl he sexually propositioned.
Original report: Man gets suspended sentence over deepfakes of 13 teenage girls
A former Scale AI employee filed a proposed class action in San Francisco Superior Court on May 9 alleging wage theft, including unpaid meeting time and no meal or rest breaks, against Scale AI and contractor HireArt, the same day the Department of Labor dropped its Fair Labor Standards Act investigation into Scale AI, HireArt and Upwork. Scale AI vows a vigorous defence.
Original report: Scale AI sued over wages as Labor Department drops probe
A Northeastern University senior found a ChatGPT prompt in lecture notes for a course whose syllabus barred unauthorized AI use, and requested a tuition refund of more than $8,000. Adjunct professor Rick Arrowood said he had run his materials through ChatGPT, Perplexity and Gamma; in May officials told her she would not get her money back.
Original report: Northeastern denies refund over professor's AI-generated course notes archived copy of Northeastern denies refund over professor's AI-generated course notes
On May 7 the Paris court ordered Orange, SFR, SFR Fibre, Free and Bouygues Télécom to block news.dayfr.com for 18 months, finding infringement established after a coalition of forty media outlets showed the site republished their articles with slight modifications. The publishers allege it uses AI to churn out up to 6,000 copied articles a day.
Original report: Paris court orders ISPs to block AI news-copying site
Youth Diversion Essex County says it receives 'two to three referrals every week' over Windsor-area students using AI to create and share fake explicit images of classmates, including a naked image of a girl that spread to multiple schools. The Windsor-Essex Catholic District School Board reported only a few such occurrences this year, with consequences typically including suspension.
Original report: Windsor students using AI nudes to bully classmates, agency says
In Fox v. Assum, Ashkelon Magistrates' Court judge Orit Salhov (Hadad) called unsatisfactory the explanation for a reference to an apparently 'imaginary' judgment, saying it was unclear how such an error happens other than through AI use, and demanded a clearer one. The court granted the defendant's request but ordered it to pay 1,200 shekels in costs, as its criticism targeted the defendant's conduct.
Original report: Israeli judge demands explanation for citation to 'imaginary' judgment
Polygon co-founder Sandeep Nailwal warned in a May 13 post on X that scammers used the compromised Telegram account of the head of Polygon Ventures to invite startup founders to Zoom calls featuring AI-generated versions of him and others. The calls, which had no audio, urged participants to install a software kit he said was designed to infect their systems.
Original report: Polygon co-founder warns of deepfake Zoom calls targeting founders
