Breakthrough ai model mythos unveiled amid cybersecurity fears

Anthropic, the AI startup behind the viral DALL-E and Claude models, has launched a top-secret initiative called Project Glasswing, partnering with Amazon, Apple, Microsoft, and Cisco to provide early access to its most advanced AI system yet, Mythos.

Mythos: a general-purpose ai for the future

Mythos, still unreleased to the public, is a general-purpose AI designed to detect security flaws in software products. Anthropic claims it discovered a 27-year-old security bug in critical internet software and a 16-year-old vulnerability in a popular video game, both missed by automated testing tools.

As cyber threats escalate, with state-backed hackers increasingly exploiting powerful AI for malicious purposes, this unprecedented collaboration reflects growing concern in the tech industry. OpenAI, a rival AI firm, has also acknowledged the rising capabilities of its own models and is working on a pilot program to put its tools in the hands of cybersecurity defenders first.

Anthropic's head of cybersecurity, Newton Cheng, emphasized that this is not just a problem for the company, but for the entire industry.