OpenAI said it banned a group of Chinese accounts, a campaign it calls Peer Review. The group used ChatGPT to debug code and write sales pitches for an AI tool apparently designed to monitor anti-Chinese sentiment on X, Facebook, YouTube and Instagram. According to OpenAI, the operators also proofread claims that their insights were sent to Chinese embassies and intelligence agents.
Haishan Yang, a University of Minnesota Ph.D. student, was expelled after a panel of four professors found his remote preliminary exam essays closely resembled ChatGPT answers. Yang denies using AI and in January filed state and federal lawsuits alleging manipulated evidence and lack of due process; according to KARE11, he said he used ChatGPT to write the filings.
Victoria Police child abuse detectives are investigating after formal photos of female students at a school in Melbourne's north-west were manipulated with AI into sexually explicit images and published online. Police say as many as 60 students could be affected. Two year 11 students have been suspended, though police have yet to speak with them. Premier Jacinta Allan called it "appalling".
Police in Attica arrested two 16-year-olds at their school on Wednesday 19 February. Police say the pair swapped photos of female classmates online after editing them with an AI program to appear nude or semi-nude, using pictures taken from social media or snapped without consent. A child pornography case file was opened against them, plus a neglect-of-supervision case against three parents. Their phones were seized.
ChatGPT disclosed the CPF identification numbers of public figures, along with company minutes and invoices filed in an electoral context, after the platform changed its personal data protection rules, Poder360 reports. OpenAI said it trains its models to refuse requests for government identifiers and is working on a fix. On 21 February 2025, ChatGPT declined Poder360's request for Bolsonaro's and Lula's CPFs.
Nudify app Clothoff claims it donates funds to "support those affected by AI" through ASU Label. Bellingcat's searches of several international non-profit databases found no registration for the group, and ASU Label declined to disclose its legal documents. ASU Label says it runs a joint deepfake study with Clothoff. Three AI detectors rated much of its own website text as likely AI-generated.
Lahore police registered three cases under the amended Prevention of Electronic Crimes Act and defamation provisions. The complaints allege that people uploaded AI-generated "immoral" videos and images of Punjab Chief Minister Maryam Nawaz and spread propaganda against state institutions. One FIR names a suspect accused of posting such a video on Facebook. Police say no arrests have been made so far.
Swiss retailer Coop confirmed to Le Temps that some branches use cameras whose footage AI analyses in real time for suspicious behaviour at self-checkouts. Coop gave no details, citing company policy, and said an entrance pictogram signals video surveillance and that no facial recognition is used. Data protection expert François Charlet said customers need clearer information than that.
On 21 February, Runjian (002929.SZ) said it had not signed a service cooperation agreement with DeepSeek. Caixin counts it as the 12th A-share company in about a week to issue such a clarification. According to Caixin, false cooperation claims posted elsewhere are being fed to DeepSeek, and its misleading answers are then spread online and fed back in.
