Getty Images' copyright lawsuit against Stability AI opened at London's High Court on Monday, with Getty alleging that Stability unlawfully scraped millions of images from its websites to train Stable Diffusion. Stability denies infringing Getty's rights and calls the case a threat to the wider generative AI industry; Getty's lawyer said the problem is AI firms using works 'without payment'.
In a June 6 letter, Senators Cory Booker, Peter Welch, Adam Schiff and Alex Padilla urged Meta to investigate and limit the 'blatant deception' of user-made Instagram AI Studio chatbots claiming to be licensed therapists, after 404 Media found bots offering fabricated license numbers and degrees. Retested, some bots refused, though a reporter-made one still supplied a 'license number'.
Córdoba prosecutors have sent to trial an adult student accused of using AI face-swapping to put female classmates' faces, girls aged 15 and 16, onto nude bodies and posting the images on porn sites with humiliating captions and links to their social media. He faces charges of serious injury aggravated by gender; the victims' lawyer says face swapping is not itself a crime in Argentina.
OpenAI says it banned several Philippines-based ChatGPT accounts behind 'Operation High Five', a campaign linked to a local commercial marketing company that generated bulk English and Taglish comments for TikTok and Facebook praising President Marcos and criticising Vice President Sara Duterte, including as 'Princess Fiona'. The network posted thousands of comments, but none OpenAI identified drew more than single-digit engagement.
Russian-language fact-checker Provereno identified more than a dozen Facebook pages posting AI-generated portraits and fabricated biographies of supposed Holocaust victims, often pairing Yad Vashem names with invented details. It says profit appears to be the motive, possibly via Meta's 'Performance Bonus' program, estimating one '90's History' post could have earned roughly $77. The Auschwitz Museum condemned the posts.
Who did it: HYBRID AI Hallucination Cases database
Chief Judge Andrew P. Gordon struck amended filings, denied a reconsideration motion with prejudice and made the National Center on Sexual Exploitation and Guinasso Law jointly liable for Las Vegas Bistro's reasonable fees after briefs cited a nonexistent case and misquoted authorities. Attorney Jason Guinasso admitted using AI; the amended motion, meant to fix the errors, still cited a case that does not exist.
Who did it: HYBRID AI Hallucination Cases database
A Mississippi federal magistrate judge recommended dismissing a prisoner's habeas petition seeking First Step Act credits for failure to exhaust administrative remedies, noting that two cases he cited, 'Gonzalez v. United States' and 'Garza v. Davis', do not exist. The judge suspected a 'growing trend' of litigants using AI to prepare briefs; Garza's citation belonged to an unrelated bankruptcy case.
Who did it: HYBRID AI Hallucination Cases database
Ordering costs against a self-represented petitioner whose Hague Convention child-return petition failed, BC Supreme Court Justice Norell found three of his cited cases bore neutral citations belonging to other, non-Convention cases, and could not locate another's stated proposition. The judge suspected internet searches and possibly AI, but accepted any errors were unintentional and did not affect the costs ruling.
Italy's data protection authority, the Garante, reaffirmed its ban on Luka Inc.'s AI companion chatbot Replika in an April 2025 decision, citing persistent GDPR violations and ongoing risks to minors and vulnerable users. Its original February 2023 order found the app lacked effective age verification, asking only for name, email address and gender.
