Anthropic's Mythos AI Sparks Alarm

The Bank of England and the U.S. Treasury are among the institutions raising alarms over the model's potential to shift the cybersecurity attack-defense balance.

A private call involving executives from Google, Microsoft, and OpenAI was convened by government officials to address these concerns.

Anthropic has withheld the most advanced version of the model, citing fears of a 'Vulnpocalypse' if such powerful tools become widely available.

This development highlights the escalating tension between rapid AI innovation and national security safeguards.