Warner Bros. Discovery sued Midjourney in California federal court, accusing it of building its business around the mass theft of content. The complaint alleges its generator returns Superman, Batman, Bugs Bunny and other characters even for prompts like 'classic comic book superhero battle'. The studio, joining Disney and Universal in suing the company, seeks its profits or $150,000 per infringed work.
Actress Angel Aquino told the Senate committee on women, chaired by Sen. Risa Hontiveros, that a friend alerted her to a deepfake porn video using her face, which she called a 'digital assault'. She urged authorities to punish those who make and share such material, the websites hosting it and platforms that 'turn a blind eye', and called for 'strong, enforceable law'.
U.S. District Judge Judith Levy dismissed Porcha Woodruff's civil rights suit against the Detroit officer who prepared the warrant for her February 2023 carjacking arrest while eight months pregnant, after facial recognition led police to put her photo in a lineup. Levy called the arrest 'troubling' but found no showing the officer lacked probable cause; her lawyer plans an appeal.
Who did it: HYBRID AI Hallucination Cases database
British Columbia's Civil Resolution Tribunal ordered X Corp. to pay a $100,000 penalty for geofencing instead of removing an intimate image, but dismissed the self-represented applicant's claim for her time after her AI-assisted reply quoted paragraph 150 of a real case only 124 paragraphs long. She had been warned about relying on AI in an earlier decision.
Guardio researcher Nati Tal says scammers are hiding malicious links in the 'From:' field of promoted video posts on X, often with adult content, which X's scans seem to miss, reaching 100,000 to over 5 million paid impressions. When users ask Grok where a video came from, the chatbot extracts the domain and repeats it in its reply.
Australia's government agreed to pay another A$475 million to victims of Robodebt, the unlawful 2015–2019 scheme that used income averaging to match welfare recipients' reported income against Australian Tax Office records and automatically issue repayment notices. The flawed system wrongly demanded money from hundreds of thousands of people; the settlement needs Federal Court approval and would be Australia's largest class-action payout.
Psychic Marcia Sensitiva said on Band's Melhor da Tarde that 30 people impersonate her online every day, including deepfake videos using her face, and that her son hired a company to track them. Host Pâmela Lucciola said Marcia and her team have taken down more than 10,000 scam sites; one case, someone claiming to sell 'soulmate portraits', ended up at DEIC.
U.S. Magistrate Judge Mark Lane recommended dismissing Jason Stanford's RICO claims as frivolous, finding his complaint cited at least two cases, including 'United States v. Green', that could not be found, and attributed quotations to two real cases that do not contain them. Lane warned Stanford, a declared vexatious litigant, of possible sanctions and ordered him to file copies of every case he cites.
Who did it: HYBRID AI Hallucination Cases database
Affirming a custody order against a self-represented father, the Iowa Court of Appeals noted in a footnote that he quoted at least one nonexistent case in a judicial-notice motion, so he 'may have misused an artificial intelligence tool'. The court denied the motion and repeated that litigants and attorneys must independently verify sources when relying on AI tools.
Who did it: HYBRID AI Hallucination Cases database
Chief Judge Stephanie M. Rose denied April Ann Nelson's motion to alter the June 30, 2025 dismissal of her suit against Navient, noting it cited nonexistent cases including 'Haglund v. Philip Morris Inc.' It was at least her third brief with such citations, the judge wrote, suggesting continued improper use of generative AI, conduct that wastes judicial resources and misleads the court.
Law enforcement agencies have quietly turned off Axon safeguards that indicate which police reports were written by AI, according to documents reported by Mother Jones journalist Tekendra Parmar. Axon's Draft One uses AI to turn body camera recordings into incident reports; many departments using it are not disclosing this, potentially leaving prosecutors and defense attorneys in the blind.
