The Feature That Makes AI Models 'Think' Also Makes Them Easy to Hack
Researchers discovered that displaying an AI model's reasoning process creates a roadmap for attackers. OpenAI's o1 rejection rate dropped from 98% to under 2%.
Tag
#deepseek
AI Privacy Audit: What Eight Popular Assistants Collect
A source-by-source audit of eight AI assistants, what they collect, how training defaults differ, and which privacy settings users can change.
Reasoning Models Hit 97% Jailbreak Success Rate
Research from ELLIS Alicante shows AI reasoning models can autonomously plan and execute attacks that bypass safety guardrails in nearly all other AI systems.
AI News Roundup: February 3, 2026
Darktrace finds 77% of security pros unprepared for AI agent threats, DeepSeek V4 imminent with coding focus, Google whistleblower alleges military AI ethics breach, and MIT warns truth verification is failing.
DeepSeek's Engram: The Memory Trick That Bypasses GPU Restrictions
China's most innovative AI lab published a technique that lets large language models store knowledge in cheap system RAM instead of expensive GPU memory. It's a direct response to US export controls -- and it works.