Australia's Administrative Review Tribunal affirmed the Privacy Commissioner's finding that Bunnings breached privacy principles on transparency and notification when rolling out facial recognition in its stores, without appropriate notice or a documented risk assessment. But it departed from the finding that Bunnings breached the rule on collecting personal information, ruling it could rely on consent exemptions to combat retail crime and protect staff and customers.
A technical note by LAPIN, Instituto Peregum and Rede Liberdade says São Paulo's Smart Sampa facial-recognition system saw at least 23 people wrongly detained over recognition errors and 82 arrested then released, and reflects racial and geographic bias. The city says all alerts are human-validated, claims 99.5% accuracy and says no unjust arrests stemmed from system alerts.
Actor Vivek Oberoi has filed a civil suit in the Delhi High Court against Collector Bazar, ZoomMantra, Indiacontent and unidentified "John Doe" defendants, alleging misuse of his identity through fake Instagram accounts, unauthorised merchandise and AI deepfakes morphing his face onto objectionable imagery. He seeks a permanent injunction and removal of the offending posters, postcards, videos and posts.
John Mearsheimer's office at the University of Chicago identified 43 YouTube channels pushing AI deepfakes of the scholar, some purporting to show him on Japan-China relations or, with a Mandarin voiceover, on fading American influence in Asia. After months of individual takedown requests, YouTube shut down 41; new channels, some calling themselves "Jhon Mearsheimer", kept appearing.
NewsGuard said seven AI-generated or manipulated images falsely linking politicians to Jeffrey Epstein drew more than 21 million views on X, including three fakes showing New York Mayor Zohran Mamdani as a child with Epstein, in which Gemini detected a SynthID watermark. Alex Jones shared one, claiming Grok had called it real.
The University Grants Commission returned dozens of PhD theses from Babasaheb Bhimrao Ambedkar Bihar University in Muzaffarpur that contained material copied from AI tools such as ChatGPT, with more than 40% similarity found, the second time it has sent theses back. Students must rewrite them and Turnitin checks are now mandatory; teachers note detection is weaker for Hindi.
Who did it: HYBRID AI Hallucination Cases database
Judge Robert Pitman, denying reconsideration of a dismissed employment suit against the University of Texas at Austin, ordered the plaintiff's counsel Gary Bledsoe to show cause why he should not be sanctioned for citing a Fifth Circuit case that apparently does not exist and mischaracterising others, which the court called hallmarks of generative AI use. A hearing is set for February 17.
Students at an Indiana intermediate school allegedly posted AI-altered photos of teachers, taken from the internet and posed with "other students" and "high profile individuals who have committed serious crimes," according to a February 3 letter to parents. The local police chief said no law prevents such actions; the school district says it handled the matter internally.
Who did it: HYBRID AI Hallucination Cases database
In a Canadian Immigration Appeal Division decision dated February 4, 2026, counsel Idorenyin E. Amana acknowledged that reply submissions were drafted with AI and not verified, after panel member Suzanne Alary found six of seven cited decisions did not appear to exist. The submissions were withdrawn; the request to reopen the sponsorship appeal was refused anyway.
Toronto lawyer Mary Hyun-Sook Lee, who told a judge a student had used ChatGPT to prepare a brief citing non-existent cases, later wrote that she used it herself and that some statements were untrue. Justice Fred Myers ordered her to show cause on criminal contempt, with a hearing due this spring, and the Law Society Tribunal has begun conduct proceedings.
