Anthropic's Mythos AI sparks security fears

The model's ability to rapidly identify and potentially exploit security holes has prompted tech giants to issue urgent software patches.

While Anthropic has opted not to release Mythos publicly, reports suggest the NSA may already be using a preview version of the technology.

This development has intensified the debate over the regulation of 'agentic' AI and the potential for a new era of automated hacking.

Experts are urging organizations to enable automatic updates to protect against exploitation by malicious actors using similar capabilities.