Anthropic is tightening the reins on its powerful artificial intelligence, citing alarming advancements in autonomous coding. The company revealed that its latest models have achieved a proficiency level that allows them to identify and exploit software vulnerabilities with a skill previously reserved for the world's top-tier security experts.

Fearing that this unprecedented capability could be weaponized by malicious actors, the firm has decided to limit public access. By imposing these restrictions, Anthropic aims to prevent the technology from falling into the wrong hands, prioritizing global cybersecurity over the rapid, unchecked release of potentially dangerous tools.