Site navigation

Research: AI Cut Vulnerability Hunt from Nine Months to 90 Hours

Elizabeth Greenberg

,

ai cyber
“AI did not eliminate the need for expert vulnerability researchers, but it dramatically compressed the work,” said Daniel dos Santos, VP of Research at Forescout.

New research from Forescout Research – Vedere Labs demonstrates how artificial intelligence is changing the economics of vulnerability discovery, dramatically reducing the time and cost required for security researchers – and attackers alike – to find vulnerabilities in foundational software commonly found in critical infrastructure environments.

The report, Finding Critical Infrastructure Bugs with Off-the-Shelf AI, examines whether security researchers and attackers can use generally available AI models and a limited budget to find vulnerabilities in both open-source software and proprietary binaries. Researchers built a specialised AI-assisted vulnerability research harness and applied it to 10 popular TCP/IP stacks – the foundational software that enables connected devices to communicate over a network – including seven previously analysed by Vedere Labs and three new stacks.

The results demonstrate how AI can significantly lower the time and cost barriers to vulnerability discovery, for both attackers and defenders. The AI-assisted approach reduced analysis time from approximately nine months to 90 hours and uncovered 99 vulnerabilities, including 66 that were previously unknown. More than half (52%) of the new vulnerabilities were rated high or critical severity, and 54 were reported to vendors and maintainers. The total cost of the automated static analysis was just $935, averaging just $2.58 per finding.

“AI did not eliminate the need for expert vulnerability researchers, but it dramatically compressed the work,” said Daniel dos Santos, VP of Research at Forescout. “With access to an effective AI-assisted workflow, vulnerabilities in complex embedded software can be found much faster and at a fraction of the cost, by security researchers and attackers alike.

“Our research demonstrates that AI is changing the economics of vulnerability discovery and accelerating the pace at which security teams need to identify, prioritise, and mitigate cyber risk. This is especially impactful for critical infrastructure, where legacy and difficult-to-patch devices may remain in service for years.”

The report found that AI cut vulnerability analysis from nine months to 90 hours. The study combined approximately 10 hours of independent AI analysis with roughly 80 hours of validation.

Researchers validated 99 vulnerabilities across 10 TCP/IP stacks, including 66 previously unknown vulnerabilities.

About half (52%) of the 66 new vulnerabilities were rated high or critical severity.

The AI-assisted system rediscovered 40% of previously known vulnerabilities, demonstrating its ability to identify flaws even in software that had already undergone extensive manual research.

Researchers reported 54 of the new vulnerabilities to vendors and maintainers. Responsible disclosure remains ongoing for some findings.

Static analysis across the 10 stacks cost approximately $935, averaging $2.58 per finding.

Binary analysis was approximately six times more expensive and took twice as long as source-code analysis, but the findings demonstrate that AI-assisted vulnerability research can extend to proprietary firmware where source code is unavailable.

Based on previous Vedere Labs research, vulnerabilities in these foundational TCP/IP components are likely to affect hundreds of downstream products from dozens of vendors.


Recommended reading


TCP/IP stacks are embedded in products used across OT, IoT, medical devices, network equipment and other cyber-physical systems. Because the same underlying software can be used across products from multiple manufacturers, a single vulnerability can create risk across the broader technology supply chain.

This supply chain risk is particularly difficult to address when vulnerable software is embedded in devices that may remain in service for decades, even after vendor support ends. Even when a fix is available, it must make its way from the software developer through firmware and product vendors before it reaches the organisation operating the affected device. For discontinued or unsupported products, a patch may never come.

“The speed of vulnerability discovery is increasing with AI, but the reality for asset owners and operators hasn’t changed,” said Barry Mainz, CEO, Forescout. “Organisations still have legacy systems, operational requirements, and devices that can’t easily be patched or simply taken offline. Containing risk, rather than patching, must be the priority. Forescout helps organisations know what is connected, understand which vulnerabilities create real operational risk in their environment, and have controls in place to contain that risk.”

Elizabeth Greenberg

Staff Writer

Latest News

Business Events Featured

Just Two Days to Go Until DIGIT Expo 2026 – Register Now!

AI Cybersecurity

Research: AI Cut Vulnerability Hunt from Nine Months to 90 Hours

Cybersecurity Editor's Picks Security

Are We Unprepared For Major Cyber Disruption?

Business

What Are The Top 4 Emerging Enterprise Risks Redefining Workforces?