Manhattan Magistrate Judge Ona Wang denied OpenAI's request to reconsider her November order to hand 20 million anonymised ChatGPT logs to The New York Times, the Daily News and other outlets suing it and Microsoft for copyright infringement. According to a lawyer for MediaNews Group and Tribune Publishing, she found OpenAI had withheld "critically important evidence"; OpenAI has appealed the original order.
Federal prosecutors have indicted Brett Michael Dadig, 31, on 14 counts of interstate stalking, cyberstalking and threats over alleged harassment of 11 women across more than five states. The indictment says he called ChatGPT his "therapist" and "best friend," and that it encouraged his podcast because "haters" meant monetization and suggested he might meet his wife at a gym.
Two Mediaset group companies, RTI and Medusa Film, have brought a case at the Civil Court of Rome against Perplexity AI. They allege it used their audiovisual and film content without permission, on a large scale, to train its AI systems. Described as Italy's first such AI copyright action, it seeks damages, a block on further use and a daily penalty.
A 17-year-old boy in Calgary has been charged with making, possessing and distributing child sexual abuse and exploitation material, plus criminal harassment. Police allege he used AI to sexualize photos of girls from several area high schools. Alberta's ALERT child exploitation unit seized two cellphones, a tablet and a laptop in a Nov. 13 search.
Kohler said its $599 Dekoda toilet camera, which photographs the bowl to advise on gut health, used 'end-to-end encryption'. Security researcher Simon Fondrie-Teitler noted it actually uses TLS, with data decrypted on Kohler's servers. Kohler says AI training uses only de-identified data from users who opt in. After TechCrunch's story, Kohler removed the claim from its website.
The Czech Constitutional Court fined a Prague lawyer 25,000 crowns over a complaint citing twelve court decisions. A substantial part of them do not exist, and part were grossly misinterpreted. The lawyer told Info.cz they had used AI under time pressure without checking it properly, and expressed regret. Info.cz says the tool was a paid professional legal system.
Hungary's Tisza Party says it is filing a criminal complaint for aggravated defamation and starting press-correction proceedings. The dispute is over posts on Viktor Orbán's Facebook page about a supposed Tisza '600-page, 1,300 billion forint austerity package'. Tisza claims Fidesz built the material with AI and, in many places, did not even remove the instructions given to the AI.
On Nov. 27, Polda Metro Jaya's cyber unit arrested a TikToker accused of making deepfakes showing Jusuf Hamka and his daughter in detainee clothing. The fakes tied them to corruption and bribery claims that Hamka's lawyers call false. Hamka says the suspect has confessed and admitted not acting alone. His lawyer calls the campaign organised and massive.
Who did it: HYBRID AI Hallucination Cases database
Chief U.S. District Judge Timothy Brooks fined lawyer Tony Pirani $1,000 over motions citing nonexistent cases and quotations. Pirani may not enter new appearances in the Western District of Arkansas until the fine is paid, must use co-counsel for two years afterwards, and will be referred to the Office of Professional Conduct. The judge said striking the AI-hallucinated filings was hardly a sanction at all.
A woman with lipedema paid $300 for six bottles of 'Svelta Venastra' skin cream. A Facebook ad appeared to show her own surgeon, along with Oprah Winfrey, Kelly Clarkson and Carnie Wilson, endorsing it, TODAY reports. The footage was digitally manipulated. The cream did not help, though her credit card company refunded her.
Ana Belén Martínez, mayor of Cúllar, has reported several AI-generated fake images to the Guardia Civil. The images, spread on social media and messaging apps, showed her in degrading, sexual and, in one case, macabre situations. The Guardia Civil's cybercrime team is investigating. The images remained visible on the Facebook profile where they first appeared.
Who did it: HYBRID AI Hallucination Cases database
The Oregon Court of Appeals struck a respondents' brief filed by Gabriel A. Watson that contained two fabricated case citations and a fabricated statement of law, and sanctioned counsel $2,000. The court said the material likely came from AI but rejected the word 'hallucination', saying the technology generates nonexistent law 'in accordance with its design'.
Who did it: HYBRID AI Hallucination Cases database
The 10th Circuit affirmed dismissal of a pro se plaintiff's discrimination suit against the City of Del City. In the alternative, it dismissed the appeal as a sanction: the brief cited eleven cases the court could not find, which it suspected were GenAI hallucinations. The plaintiff's show-cause response cited another nonexistent case. Future filings must disclose any GenAI use under penalty of perjury.
