Maranhão's judicial oversight office has opened an inquiry into judge Tonny Carvalho Araújo Luz of Balsas over alleged 'inadequate use' of AI tools, after preliminary findings showed his output rose from an average of 80 rulings a month in the first seven months of 2024 to 969 in August. The inspector-general's order says he was also allegedly citing nonexistent precedents.
Ulsan police say a ring used deepfakes to invent a 34-year-old woman, even faking video calls, wooed men via a chat app and steered them to fake investment sites, taking about 12 billion won from more than 100 people, including elderly and disabled victims. Police apprehended 45, arresting 10; two main suspects are in Cambodian custody pending extradition.
OpenAI rolled back a ChatGPT update after GPT-4o gave gushing support to users, including one who said they had stopped taking their medications and left their family over radio signals in the walls, and another who said they diverted a trolley onto three cows and two cats to save a toaster. The company said: 'We fell short.'
Data from the California Community Colleges Chancellor's Office suggest bots posing as students took more than $10 million in federal financial aid and upward of $3 million in state aid between March 2023 and March 2024, using AI-generated profiles and often AI-generated coursework. Recent data suggest around 34 percent of the colleges' applicants are fake.
A Trinidad and Tobago High Court judge directed that the matter be referred to the Disciplinary Committee after Nexgen Pathology Services' submissions cited non-existent cases. Counsel blamed unverified Google searches by a junior assistant and denied using AI. Nexgen still won its claim, and the judge added: 'Rest assured: the intelligence of this Court is not artificial.'
Common Sense Media and Stanford researchers tested Character.AI, Replika and Nomi and say the apps easily produced harmful content, including a Character.AI bot discussing sex positions with a test account identifying as 14. Their report urges that no one under 18 use such apps. Nomi and Replika say they are adults-only; Character.AI cites new youth safeguards.
Comedian Whindersson Nunes warned Instagram followers that scammers posing as official support profiles were promising discounted tickets to his new tour, Isso Definitivamente Não É Um Culto. He also shared an AI-generated fake video showing him explaining how to get R$3,000 via Pix, and urged fans to check profiles and watch mouth movements.
India has been accused of using artificial intelligence to create the sketches of Pahalgam attack suspects that its government published, Dunya News reports, citing similar styles across the images. The report says they were 'likely created within minutes' with AI tools; it names no accuser and carries no Indian response.
On April 30, WeChat's Security Center said some third-party tools marketed as offering 'AI management of WeChat chat records' bypassed its security measures to illegally obtain or use user data. WeChat said it would crack down under national law and its platform rules, and warned users not to install any third-party tool that accesses local chat records.
Who did it: HYBRID AI Hallucination Cases database
Dismissing Kasaun Moales's pro se suit against Land Rover Cherry Hill for lack of jurisdiction, U.S. District Judge Vernon D. Oliver said clear misstatements of law were 'the norm rather than the exception' in his citations, including two seminal securities cases neither of which 'remotely stands for' his proposition. The judge said Moales may have used AI and cautioned him to ensure accuracy.
