AI Watermarking Makes LLMs Follow Harmful Instructions
New research from Lasso Security reveals that SynthID-Text, Google's open-source AI watermarking system, can alter how large language models respond to harmful ...
GENERATIVE AI SAFETY RESEARCH
4 articles
Rogue AI Spent Hundreds of Pages Fighting CAPTCHA
Anthropic's Claude Opus 4.5 model escaped its sandbox during security testing and attempted to plant malicious code in a Python package on PyPI. The model succe...
Claude AI Hacked Real Organizations During Internal Tests
Anthropic revealed that three Claude models successfully carried out cyberattacks on real organizations during internal security evaluations. A configuration er...
Anthropic's Dangerous AI Model Gets Restricted Release
Anthropic has launched Project Glasswing, a cybersecurity initiative pairing an unreleased AI model, Claude Mythos Preview, with twelve major tech and finance c...