The daily rap sheet of the machine age.
13 items
Disney and Universal have sued Midjourney in federal court in Los Angeles, alleging its image generator is a "bottomless pit of plagiarism" that makes "innumerable" copies of characters such as Darth Vader, Elsa and the Minions; the complaint says Midjourney made $300m last year alone. Midjourney did not immediately respond to the BBC's request for comment.
Original report: Disney and Universal sue Midjourney over copied characters
Google's AI Overviews told a searcher on Reddit that the Air India Flight 171 crash, in which more than 200 passengers and crew reportedly died, involved an Airbus A330; it was a Boeing 787. Ars's own searches returned Boeing, Airbus or a mashup of both. Google says it manually removed the response.
Original report: AI Overviews swaps Boeing for Airbus in Air India crash
Meta says it has taken legal action against Joy Timeline HK Limited, following what it describes as multiple attempts by the company to circumvent its ad review process and keep placing ads for AI "nudify" apps after they were repeatedly removed. Meta also says it has shared more than 3,800 URLs of such apps and sites with other tech companies since the end of March.
Original report: Meta takes legal action against nudify app advertiser
Ireland's Data Protection Commission fined the Department of Social Protection €550,000 and issued a reprimand over facial matching used to register people for the Public Services Card, finding GDPR breaches including a failure to identify a valid lawful basis for collecting biometric data. The department, which says it has one, has nine months to identify one or stop the processing.
Original report: Data watchdog fines Irish department €550,000 over facial matching
Around 10,000 Model 3 and Model Y owners have reportedly signed up to a class action against Tesla Australia, which alleges the cars' automatic emergency braking engages abruptly in inappropriate circumstances, that range falls short of advertised figures, and that the hardware cannot support autonomous driving despite the Autopilot label. A verdict is still forthcoming.
Original report: About 10,000 Tesla owners join Australian phantom-braking class action
"Divorce Court" judge Faith Jenkins and her husband, R&B singer Kenny Lattimore, have filed a lawsuit against anonymous YouTubers they say spread lies about their marriage. The videos reportedly used AI-generated content, fake court documents and fabricated quotes; Jenkins wrote that no one should profit "from defamation while hiding behind anonymity."
Original report: Faith Jenkins and Kenny Lattimore sue anonymous YouTubers over AI videos
Tesla has sued former engineer Jay Li and his startup Proception in California federal court, alleging Li downloaded confidential files on Optimus robotic hand research before leaving and used them to shortcut development. The complaint says Li founded Proception six days after leaving Tesla and that its hands bear "striking similarities" to Tesla's.
Original report: Tesla sues former Optimus engineer over robot hand secrets
TechCrunch reports that some users of the stand-alone Meta AI app appear unaware they are publicly sharing conversations, audio and images, including questions about tax evasion and a reference letter naming an employee facing legal trouble; security expert Rachel Tobac found home addresses and court details. Meta does not show users their privacy settings as they post.
Original report: Meta AI app users appear to publish private chats unawares
A deepfake ad uses musician Vladi Ampov, known as Grafa, to promote a program promising a sixfold increase in viewers' bank balances if they register on a website and pay a small sum. Built from real footage with his words replaced, it reached him via a friend; he and his team reported it.
Original report: Deepfake Grafa promises riches from the sofa
A TikTok profile has posted videos, made with Google's Veo 3, depicting a nationwide strike by truckers, motorcycle couriers and agribusiness workers in support of Jair Bolsonaro, according to fact-checkers Comprova; no such mobilisation occurred. The clips carry a Veo watermark, Google's SynthID flagged them as synthetic, and by June 11 they had 925,000 combined views.
Original report: AI-made TikTok videos fake trucker strikes backing Bolsonaro
Researchers from Stanford, Cornell and West Virginia University estimate that Meta's Llama 3.1 70B has memorized 42 percent of Harry Potter and the Sorcerer's Stone well enough to reproduce 50-token excerpts at least half the time, against 4.4 percent for Llama 1 65B. It memorized only 0.13 percent of Sandman Slim, by Richard Kadrey, lead plaintiff in a class action against Meta.
Original report: Llama 3.1 70B recalls 42% of first Harry Potter book
A Massachusetts mother filed a Title IX complaint after, she says, another student used AI to create a pornographic deepfake image of her middle-school daughter. After about four and a half months, the district called the conduct "inappropriate and hurtful" but found insufficient evidence it occurred in its schools; she says the image was shared in school hallways.
Original report: District declines to discipline student over deepfake nude image
A video circulating on TikTok from June 11, 2025 appeared to show 'El Gordo y la Flaca' host Lili Estefan announcing that Mexican actress Adela Noriega had died of cancer. Pulzo reports the news was false and apparently made with AI, combining manipulated audio with old footage; the show's real broadcast that day showed different clothing and a different set.
Original report: Apparent AI video falsely announces actress Adela Noriega's death
