Cybersecurity
China’s Open Model Is Now the World’s Cybersecurity Stress Test
The AI Prism Admin
Z.ai's open-weight GLM-5.3 has found 2,436 verified vulnerabilities, edged past Anthropic's Mythos 5 on one cyber benchmark, and arrived in the middle of a summer when OpenAI, Anthropic, Meta and Moonshot all lost control of their own agents. The NYT calls it a test of the world's cybersecurity — and the test is already running.
The AI Worm Is Already Here — It’s Crawling Through Copilot for Word
The AI Prism Admin
A security researcher has demonstrated an AI worm that self-propagates through Copilot for Word, altering documents and spreading hidden instructions across your organization. Here's how it works, why Microsoft can't patch it, and what you can do.
