AIMIM president Asaduddin Owaisi complained to Hyderabad police about a 7.35-minute AI-generated video in which he appears to promote a platform promising Rs 53,000 a day, alongside AI-generated Nirmala Sitaraman, Mukesh Ambani, Narayana Murthy and Amitabh Bachchan. Cyber Crime police registered a case under IT Act and BNS sections including identity theft.
OpenAI's latest report on misuse of its models, released Thursday, outlines 10 cases, including ChatGPT accounts of probable Chinese origin posting and commenting on each other about Taiwan and USAID to create an illusion of authentic engagement, and another China-linked case involving password bruteforcing. Other cases were allegedly linked to Russia, Iran and Cambodia; China's foreign ministry denied involvement, according to Reuters.
A school in Minas Gerais told students accused of using AI to make fake nude images of female classmates to stay away from June 4 as a precaution, and reported the case to child-protection authorities and prosecutors. According to complaints, students had made such images for over a year and shared them in school groups and Telegram forums; civil police have opened an inquiry.
London's High Court on Friday lambasted lawyers in two cases who apparently used AI tools to prepare written arguments citing fake case law. Judge Victoria Sharp warned that citing non-existent cases breaches the duty not to mislead the court and could amount to contempt, or, in the most egregious cases, the criminal offence of perverting the course of justice.
A DOGE engineer with no health care or government experience built an AI tool that flagged more than 2,000 Veterans Affairs contracts as "munchable," ProPublica reports, misreading values so that more than a thousand were deemed worth $34 million each, some actually worth as little as $35,000. Engineer Sahil Lavingia concedes "mistakes were made" but says no contract was terminated without human review.
Former Corinth School District superintendent Edward Childress has been indicted on federal and state charges, as has ex-teacher Wilson Frederick Jones, accused of using AI to turn students' photos into explicit material. Court documents say Childress knew in November but told state education officials only in January; the indictment alleges he let Jones resign and misrepresented why to the school board.
Meta's Oversight Board says the company is probably allowing significant scam content to avoid over-enforcing against a small subset of genuine celebrity endorsements, citing a Plinko casino-style game ad using an AI-manipulated video of Ronaldo Nazário that was reported over 50 times and removed only after the board took the case, by then past 600,000 views. Meta calls many of the claims "simply inaccurate."
A CBS News investigation found at least hundreds of ads for "nudify" apps, AI tools that make sexually explicit deepfakes of real people, across Facebook, Instagram, Threads and Messenger, some promising to "see anyone naked." Meta says it removed the ads and deleted the Pages responsible, but CBS News found such ads still running on Instagram afterwards.
A civil court in Niterói ordered Google to pay R$20,000 in moral damages after its Gemini AI wrongly named a Brazilian entrepreneur in search results as owner of an online betting site, which court documents say led to harassment and death threats. The plaintiff's lawyer says Gemini misread an interview recommending seven video games; Google has not appealed.
An AFP investigation found dozens of YouTube and TikTok pages churning out hundreds of AI-generated sermons and speeches attributed to Pope Leo XIV, with AI disclaimers often hard to find and sometimes absent. After AFP flagged them, YouTube terminated 16 of 26 channels plus one more, and TikTok removed 11 accounts; one TikTok fake, despite an AI label, drew some 32.9 million views.
Poland's data protection office, UODO, says peers used an AI app to alter an image of a girl under 15, with reports of attempts to share it, most likely with pupils at the same school. The office says this may amount to producing or spreading pornographic content involving a minor, punishable by two to 15 years in prison.
Police and the Bucks County District Attorney's Office in Pennsylvania are investigating allegations that a middle school student used AI to create pornographic deepfakes of classmates, with photos of as many as 12 underage girls used, according to people familiar with the allegations. If confirmed, the case could be among the first prosecuted under a new state law outlawing their distribution.
Who did it: HYBRID AI Hallucination Cases database
Paraná's state court fined an appellant 1% of the case value for bad-faith litigation after the appellant's lawyer cited nonexistent precedents. The judge grew suspicious when one fit the case impressively well, its real REsp number belonging to an unrelated case, and the lawyer admitted not checking a filing prepared with an "AI search tool."
