AI models find dozens of unknown flaws in network software, including on unsupported devices
Forescout researchers tested whether widely available AI models could find security flaws on a limited budget, in both open-source and compiled code. They confirmed 99 vulnerabilities across 10 TCP/IP stacks. Sixty-six were previously unknown, and over half of those were rated high severity or worse.
Why it matters
Attackers can use the same cheap method, so organizations running old devices that no vendor will patch face growing risk and should isolate or replace them.
aii.news picks and summarizes stories from trusted news sources with the help of AI. Every story links to the original reporting. How we work.