Too risky: Why Anthropic is not releasing its new cyber AI to the public

[Redaktion]

Anthropic is keeping its new cyber AI, Claude Mythos, under wraps, even though the company says the model has found security vulnerabilities in major operating systems and browsers. The reason: Mythos may not only be able to detect vulnerabilities, but also help create dangerous exploits.

https://www.notebookcheck.net/Too-risky-Why-Anthropic-is-not-releasing-its-new-cyber-AI-to-the-public.1268816.0.html