Should the Mythos AI model raise cybersecurity alarms?

On April 7, AI steadfast Anthropic said its caller model, Mythos, is truthful almighty astatine uncovering cybersecurity vulnerabilities that it volition not beryllium released publicly. Instead, the institution said it volition beryllium shared with firms that physique captious bundle utilized crossed the economy. The initiative, known arsenic Project Glasswing, has raised cybersecurity concerns among astir firms, which indispensable present contend with AI-enabled attackers utilizing alleged “zero-day” exploits, which are chartless adjacent to a fixed portion of software’s developers, and truthful are not patched done bundle updates. The Hindu reported past week that the Union authorities and the Indian IT sector’s main cybersecurity assemblage are some studying the implications of Mythos. Should the Mythos AI exemplary rise cybersecurity alarms? Aseem Jakhar and Sharda Tickoo sermon this successful a speech moderated by Aroon Deep.

Aseem Jakhar is the laminitis of Payatu and co-founder of the Nullcon cybersecurity conference; Sharda Tickoo is the state manager, India astatine Trend AI, formerly Trend Micro