Forescout says AI cuts vulnerability discovery time from nine months to 90 hours
AI Cuts Vulnerability Discovery from Nine Months to 90 Hours
Forescout used its own AI-assisted vulnerability research tool to find 99 vulnerabilities in 10 commonly used TCP/IP protocol stacks, 66 of which were zero-days, compressing analysis time from about nine months to 90 hours. The team then spent another approximately 80 hours manually verifying these findings. The report states that static analysis of the 10 protocol stacks cost about $935, equivalent to $2.58 per finding; binary analysis cost about six times as much as source code analysis and took twice as long. Forescout research vice president Daniel dos Santos said AI has not replaced specialized vulnerability researchers but has greatly compressed the workload. The article also recommends that security teams prioritize risks by reachability, exploitability and operational impact, shut down unnecessary service exposure such as FTP, Telnet, IPv6 and web management interfaces, isolate devices that cannot be patched, accelerate patch cadence, and in procurement require suppliers to clearly state support cycles, publish vulnerability disclosure policies, provide SBOMs and continuously conduct security testing.