Open-Weight AI Models Are Closing the Capability Gap—But the Safety Divide Widens⭐8

Open-weight AI models like China's GLM-5.2 peer with top proprietary systems in capability, yet safety refusals lag sharply, widening a critical risk gap that r...

The White House Wants Anthropic to Block All Jailbreaks. That May Not Be Possible⭐8

The White House demands Anthropic block all jailbreaks for Fable 5, but AI safety experts say perfect prevention may be impossible.