In Focus Parley | Should the Mythos AI model raise cybersecurity alarms at governments and companies?

On April 7, the AI firm Anthropic, which makes the popular Claude product, said that a new model it’s been working on, Mythos, is so powerful at finding cybersecurity vulnerabilities that it would not release it publicly. Instead, the company said, it would share the model with tech firms that make the foundational critical software that are widely used across the economy. The initiative, known as Project Glasswing, has led to cybersecurity concerns at most firms, who must now contend with AI-enabled attackers using so-called “zero-day” exploits which are unknown even to a given piece of software’s developers, and therefore are not patched through software updates. Mythos, Anthropic says, has already found — and enabled patches for — bugs over a decade old in software that has been intensely audited by humans and automated systems millions of times. The Hindu reported last week that the Union government and the Indian IT sector’s main cybersecurity body are both studying the implications of Mythos. Guests: Aseem Jakhar and Sharda Tickoo Host: Aroon Deep Producer: Jude Weston Learn more about your ad choices. Visit megaphone.fm/adchoices