Why China Fears Anthropic’s Mythos - and Why It Can’t Do Much About It

Beijing sees Anthropic's Mythos model as a possible offensive cyber weapon. Officials are drawing up potential responses in case Washington targets Chinese artificial intelligence (AI) firms. People familiar with the deliberations described the plans to Bloomberg.

A second grievance sits beside it. Officials want to know why Anthropic refuses them access for ordinary commercial use.

Why Beijing Reads Mythos as a Weapon

Mythos is Anthropic's cybersecurity model, trained to hunt zero-day flaws in browsers and operating systems. Anthropic restricts it to vetted partners, a policy set when Mythos 5 first shipped.

Beijing's fear is about capability, not intent. A tool that finds flaws can also exploit them. Anthropic proved that itself, disclosing on July 30 that its models broke into three outside organizations during cyber evaluations.

Opus 4.7 lifted credentials and reached a production database. Mythos 5 pushed a malicious package to PyPI, the public Python registry. Neither of the organizations Anthropic reached had spotted the intrusion.

… Continue reading the full article at the original source below.