US District Judge Beth Bloom on February 19, 2026 denied Tesla's bid for judgment as a matter of law or a new trial, leaving intact a $242,570,000 judgment, including $200,000,000 in punitive damages, over a 2019 Key Largo crash involving an Autopilot-equipped Model S that killed one person. The order calls Tesla's grounds 'virtually the same' as those already rejected.
Amazon says a brief service interruption last December, which the Financial Times attributed to AI, was 'user error—specifically misconfigured access controls'. According to Amazon, it hit only AWS Cost Explorer in one of 39 regions, drew no customer inquiries, and prompted safeguards including mandatory peer review for production access; it calls the FT's report of a second event 'entirely false'.
Microsoft deleted a November 2024 blog post, written by a senior product manager, that pointed developers to a Kaggle dataset of all seven Harry Potter books, wrongly marked public domain, to build AI Q&A and fan-fiction tools, after a Hacker News backlash. One example story had a new friend on the Hogwarts Express sell Harry on Microsoft's Native Vector Support in SQL.
Publisher Finji says TikTok ran AI-altered ads for its games without permission despite the company turning off genAI options, including one, according to IGN, showing Usual June's Black protagonist with a bikini bottom and 'impossibly large hips and thighs'. TikTok support first claimed Finji uploaded the image itself, then cited a 'catalog ads format' with no guaranteed opt-out.
Italy's privacy regulator, the Garante, fined Università eCampus €50,000 for unlawfully processing biometric data of participants in online teacher-qualification courses, using facial recognition to verify identity and attendance. The authority found no suitable legal basis and no prior impact assessment; more than 450 participants per lesson were affected. It credited the university's cooperation and voluntary halt of the system.
The Delhi High Court on Friday granted actor Kajol interim protection of her personality rights, with Justice Jyoti Singh restraining several defendants from commercially using her name, image, voice or likeness and directing removal of pornographic and obscene material published online against her. The court said her rights must be protected from misuse through artificial intelligence and deepfake technology.
The Bank of Greece says social media posts showing Yannis Stournaras chatting with journalists and steering viewers to an online 'investment' platform, promising very large profits for a small stake, were made with AI tools to defraud the public. The central bank notes it takes no deposits and urges citizens not to send money or share account or card details.
Police in Touba, acting on a complaint filed February 16, 2026, arrested a man accused of using a woman's WhatsApp status photos and videos to build fake profiles, then altering his voice with AI tools on calls to pose as a woman, sending sexual content and soliciting money or phone credit. He confessed, police say, and remains in custody.
