X's safety team says it found a bot farm of about 200,000 inauthentic Chinese accounts, 200 of which posted in a way that could manipulate US debate on AI and energy policy, claiming data centers raise household electricity prices; some accounts shared AI-generated cartoons of developers with bags of cash. OpenAI earlier banned a similar ChatGPT account cluster it said likely originated in China.
Spain's Policía Nacional in Langreo is investigating fake nude images made with AI tools from photos young women had posted on social media and then shared in a WhatsApp group. Police say 15 possible victims have been identified, 13 have filed complaints, and a suspected perpetrator has been identified and reported to the judicial authority.
Justice André Mendonça of Brazil's Superior Electoral Court gave X 24 hours to explain its mechanism for keeping candidates' posts out of the "For You" tab, after the Brasil Pronto Pra Mais coalition said X's published code covered only 665 candidate accounts and missed 1,559 others, which it argued left some candidacies amplified.
US ambassador Bill White posted an AI-generated image of Belgian health minister Frank Vandenbroucke on Instagram, captioned "Harms all forms of joy in life" and "Who is the biggest loser in Belgium?", amid a row over stricter alcohol warning labels. Foreign Minister Maxime Prévot called this unacceptable and said the ambassador would be summoned, according to Belga.
Broadcaster Lee Ji-hye warned followers on the 28th about an AI-made video ad that shows a synthesised Lee eating and promoting a product, with a sales link attached. Lee said the seller seemed not to be a domestic company and that legal options were being explored; Lee flagged similar fake underwear and food ads in June.
A Seventh Circuit panel upheld dismissal of the home-possession charge against Steven Anderegg, who prosecutors say used Stable Diffusion to make more than 13,000 sexually explicit images of prepubescent children, finding the First Amendment protects private in-home possession of material depicting no real child. Production, distribution and transfer-to-a-minor charges remain pending. Two judges asked the Supreme Court for guidance.
A federal judge in Atlanta sanctioned plaintiff's counsel Mr. Johnson in Booker v. Kroger, finding he cited fake or hallucinated cases, told the court at a hearing "I did not use artificial intelligence", then admitted he had "consulted" AI tools. The court found he repeatedly lied and fined him $1,000 for each of four fake or misleading authorities.
After the Bhotekoshi flash floods, Nepal's government asked Meta and TikTok to remove graphic images, AI-synthesised videos passed off as disaster footage, and fraudulent donation QR codes faster than the 24 to 48 hours they currently take, according to NTA director Bijay Kumar Roy. Nepal Police are identifying suspected creators of malicious content.
