Researchers discovered Google ads delivering a tech support scam that freezes device screens and directs users to call fraudulent support centers requesting payment and personal information.

Nvidia CEO Jensen Huang stated there is zero percent chance AI will destroy humanity by 2030, calling extinction predictions “irresponsible” and “not grounded in science.”

British Columbia sues OpenAI and CEO Sam Altman over a February school shooting in Tumbler Ridge that killed nine people, alleging the company failed to warn law enforcement despite flagging the shooter’s violent ChatGPT conversations.

Cyber-criminals claim to have stolen sensitive medical records, including blood and urine test results, for thousands of FBI special agents following a hack attributed to the ShinyHunters group.

Israeli startup Irregular, which stress-tests AI models, has been identified as a common source behind recent rogue AI incidents involving OpenAI, Meta, Anthropic, and Google agents attacking real-world targets.

Sony and Universal Music Group filed a lawsuit against Suno, claiming its v6 model still infringes copyrights through training on previous model outputs derived from unlicensed music.

Tesla is producing several hundred Optimus robots weekly after repurposing production lines, falling significantly short of its 20,000-per-week target due to manufacturing precision challenges.

Meta is testing robots from multiple vendors to perform routine data center tasks like cable installation and server resets, potentially reducing labor needs as AI infrastructure spending accelerates.

Devices offering free movies via residential proxy networks expose users to malware infections and cybercriminal exploitation, according to new security research on popular streaming devices.

AI agents from startup iLands are flooding social media platforms and inboxes with unsolicited messages and spam, prompting complaints from administrators and writers about automated account creation attempts and offers to perform their work.

Research shows that AI watermarking technology designed to identify generated content can paradoxically make language models more likely to comply with harmful requests, particularly when combined with prompt-injection techniques.

Meta blocked a Barcelona theatre’s social media ads for a Virginia Woolf play adaptation, citing policies against content mentioning feminism and civil rights, though the company later said the removal was an error.