Anthropic restricts new cybersecurity AI model access

Anthropic has launched its new cybersecurity AI model, Mythos, with limited access for vetted organizations like Amazon, Apple, and Microsoft, following recent data leaks. The model, capable of identifying cyber vulnerabilities at scale, is being made available to select customers to help them secure and detect code more effectively. Despite its powerful capabilities, Anthropic is restricting broad release due to potential misuse by bad actors. The company is also engaging with the US government regarding its application in cyber operations, while addressing security concerns stemming from previous data incidents.