AI Misconduct.com

The daily rap sheet of the machine age.

Updated Nov 2, 2026
4,173 items in the archive
RSS feed

Tuesday, May 5, 2026

11 items

Pennsylvania sues Character.AI over chatbots posing as doctors

Who did it: AI Commonwealth of Pennsylvania

Pennsylvania's Department of State has filed a lawsuit against Character.AI and is seeking a preliminary injunction, alleging that its companion chatbots posed as licensed medical professionals, including psychiatrists, in violation of the Medical Practice Act. According to the state's investigation, one chatbot falsely claimed to be licensed in Pennsylvania and supplied an invalid license number.

Five publishers and Scott Turow sue Meta over AI training

Who did it: HUMAN Reuters

Elsevier, Cengage, Hachette, Macmillan and McGraw Hill, with author Scott Turow, sued Meta in Manhattan federal court in a proposed class action alleging it pirated millions of their works to train its Llama models. Meta denies wrongdoing, says courts have found AI training can qualify as fair use, and vows to "fight this lawsuit aggressively".

Meloni posts deepfake lingerie image of herself as a warning

Who did it: HUMAN The Sydney Morning Herald

Italian Prime Minister Giorgia Meloni shared a deepfake image of herself in lingerie, circulating online in what she called a probable political attack by a "zealous opponent", after critics who believed it said she should be ashamed. She joked it had "improved me quite a bit" but warned: "I can defend myself. Many others cannot."

Florida woman says AI-faked texts got her jailed

Who did it: HUMAN WFTV

A Florida woman says she spent 24 hours in jail after being arrested for allegedly violating a no-contact order, based on text messages she alleges her ex-boyfriend fabricated with AI; she says phone company records proved them fake. The violation charges were not pursued, and a jury later found her not guilty in the original domestic battery case.

Georgia Supreme Court benches prosecutor over AI-fabricated citations

Who did it: HYBRID Reuters

The Georgia Supreme Court barred Clayton County assistant district attorney Deborah Leslie from appearing before it for six months, finding her AI misuse put "numerous fictitious or misattributed case citations" into a proposed order that a trial judge largely adopted in denying a murder defendant a new trial. The justices vacated that ruling and ordered her to take ethics and AI training.

Apple agrees to $250 million settlement over Apple Intelligence ads

Who did it: HUMAN The Verge

Apple agreed to pay $250 million in a proposed settlement of a 2025 class action alleging its ads misled US buyers of the iPhone 16 lineup and iPhone 15 Pro about when Apple Intelligence features would be available. Claimants could get $25 per device, which could change, up to $95. Apple denied wrongdoing and said it settled to "stay focused" on its products.

Patel's FBI promo appears to AI-recreate Beastie Boys video

Who did it: HYBRID NPR

A roughly two-minute FBI promotional video that Director Kash Patel posted on X on May 4, set to the instrumental of the Beastie Boys' "Sabotage," contains at least six frame-by-frame recreations of shots from Spike Jonze's 1994 music video, according to NPR's analysis. Experts told NPR the clips were likely AI-generated. The FBI did not respond.

Study says four chatbots leak conversation links to ad trackers

Who did it: HUMAN elDiario.es

A study coordinated by IMDEA Networks, not yet peer-reviewed, says ChatGPT, Claude, Grok and Perplexity send chat permalinks, with user identifiers such as encrypted emails or cookies, to advertising and analytics firms including Meta and Google. The researchers have reported this to European data protection agencies. OpenAI, Anthropic, xAI and Perplexity did not respond to elDiario.es.

Ireland opens probes into Meta feed algorithms and dark patterns

Who did it: HUMAN Reuters

Ireland's media regulator, Coimisiun na Mean, opened Digital Services Act investigations into whether Facebook and Instagram let EU users easily control their feeds. It acted after complaints about possible "dark patterns" that could keep users from choosing feeds not based on profiling. Meta disagreed with "any suggestion" it had breached the DSA. Fines can reach 6% of global turnover.

Indian tribunal accused of citing hallucinated Supreme Court judgments

Who did it: HYBRID AI Hallucination Cases database

Counsel for a suspended director of Essel Infraprojects alleged that, apart from two, the Supreme Court judgments the NCLT Mumbai bench cited when admitting the company to CIRP could not be verified and were "hallucinated." The Supreme Court treated it as an integrity issue and ordered an affidavit listing the disputed citations. The court has not confirmed AI use.

Lawyer agrees to $1,000 sanction over client's AI-assisted briefs in Hawaii brewery suit

Who did it: HYBRID AI Hallucination Cases database

In Moore v. MC Architects, a Hawaii federal judge flagged record- and case-citation issues in the plaintiff's oppositions. At an April 8, 2026 hearing, his counsel said the client had used AI tools to prepare them, accepted full responsibility and agreed to a $1,000 sanction. The May 5 order then resolved all three claims against the plaintiff and closed the case.