TikTok personality Alina Amir says a video circulating as her 'leaked video' was AI-generated. She stayed silent for a week, until hundreds of posts claimed her video had leaked. She then announced a cash reward for credible information identifying its creator and asked Punjab Chief Minister Maryam Nawaz Sharif to act against people who make AI videos of women.
The European Commission said on January 26 it would investigate under the Digital Services Act whether X properly assessed and mitigated risks from Grok, after the chatbot produced sexualised images of women and minors. X pointed to xAI's January 14 restrictions on image editing, which a senior Commission official said were welcome but did not resolve all the issues.
A class action filed Jan. 23 in the Northern District of California alleges xAI failed to implement safeguards against Grok creating thousands of non-consensual deepfake images on X. According to the complaint, Grok turned the anonymous lead plaintiff's fully clothed photo into a revealing bikini image. Asked for comment, xAI sent an auto-generated email: "Legacy Media Lies."
Google has agreed to pay $68m to settle a class action claiming Google Assistant sometimes switched on by mistake and recorded conversations meant to be private. The proposed settlement, filed in a California federal court, needs approval from US District Judge Beth Labson Freeman; Google denies wrongdoing. Plaintiffs' lawyers may seek up to a third, about $22m.
Creators behind three YouTube channels, led by h3h3 with golf channels MrShortGame Golf and Golfholics, filed a proposed class action in the Central District of California alleging Snap trained AI features such as Imagine Lens on their videos via the research-only HD-VILA-100M dataset, circumventing YouTube's restrictions. They have already sued Nvidia, Meta and ByteDance. Snap declined to comment.
A lawsuit by the State of Illinois and the City of Chicago alleges DHS agents use the Mobile Fortify app to run facial recognition on minors who are US citizens, without consent or individualized suspicion, and keep the data for up to fifteen years. The complaint says the app has been used in the field more than 100,000 times since around June 2025.
Who did it: HYBRID AI Hallucination Cases database
Judge Kearney of the Eastern District of Pennsylvania on January 26, 2026 sanctioned New York lawyer Yen-Yi Anderson $4,000 after a motion to dismiss in Lifetime Well v. IBSpot.com contained at least eight false citations; a law clerk had used Lexis+ AI and LexisNexis Protégé. Anderson fired the clerk; local counsel Jeffrey Goldin, who signed without reading the cited cases, got a non-monetary sanction.
Justice Ravi Nath Tilhari of the Andhra Pradesh High Court found a trial court order cited non-existent, AI-generated citations, which the judicial officer admitted using in good faith, but held the order stands because the law it applied was correct. The court warned AI may "fabricate authorities" and said judges must independently verify all authorities they rely on.
Who did it: HYBRID AI Hallucination Cases database
Judge Jon S. Tigar of the Northern District of California on January 26, 2026 declined to sanction pro se defendant Tony Chou, whose opposition in a copyright suit by MacroCharts Research cited what the court called nonexistent or "hallucinated" authorities. Chou had filed an errata removing them; the court called the filing irresponsible and ordered future citations verified.
Who did it: HYBRID AI Hallucination Cases database
A Middle District of Florida judge on January 26, 2026 dismissed without prejudice Ibrahim Shehatta Gerges Farag's $20,000,000 suit against a probate lawyer and law firm, noting it cited two cases that do not exist, misquoted a Florida statute and misdescribed its exhibits. The court warned that if generative AI was used, further fake citations could bring Rule 11 sanctions.
