The AI model release pace is accelerating so quickly that independent evaluation is struggling to keep up. New releases from Meta, xAI and Z.AI arrived within days of one another this month, while recent launches from OpenAI and Anthropic show how quickly the competitive frontier is moving. AI model release pace keeps accelerating The broad

AI agent security is entering a tougher phase as companies move from experiments to production. Surveys now point to widespread agent breaches, access risk and data exposure, while researchers and AI developers have documented cases in which autonomous systems ignored instructions or crossed intended boundaries. Agent breaches expose enterprise risk Research associated with the Cloud Security A

The AI chip cloud loophole has become a major challenge for US export policy after officials accused Moonshot AI of accessing restricted Nvidia GB300 systems outside China. Kimi K3’s strong performance has intensified debate over remote chip access, export controls and the Remote Access Security Act. AI chip cloud loophole US semiconductor restrictions were largely built around controlling ex

ChatGPT for Teens is rolling out with stronger teen AI safety measures, parental controls and learning-focused features. OpenAI says eligible under-18 accounts can automatically enter the experience, which combines age prediction, Study Mode, Quiet Hours and additional safeguards for sensitive content. ChatGPT for Teens OpenAI has started rolling out ChatGPT for Teens, a dedicated experience de