AI cuts networking flaw hunting to 90 hours: Forescout
Thu, 19th Nov 2026 (Today)
Forescout has published research showing that artificial intelligence cut vulnerability hunting in networking software from about nine months to 90 hours. The study identified 99 vulnerabilities in software used by connected devices in critical infrastructure environments.
Vedere Labs researchers tested 10 popular TCP/IP stacks, the software components that allow devices to communicate over networks. These stacks are commonly embedded in operational technology, internet of things devices, medical equipment, and network products.
Of the 99 validated vulnerabilities, 66 were previously unknown. Of those new flaws, 52% were rated high or critical severity, while 54 were reported to vendors and maintainers.
The research examined whether generally available AI models could help researchers and attackers find weaknesses in both open-source software and proprietary binaries on a limited budget. To test this, the team built an AI-assisted vulnerability research harness and applied it to software that, in some cases, had already been examined in earlier manual research.
The AI-assisted system rediscovered 40% of previously known vulnerabilities in the selected TCP/IP stacks. That suggests the method can identify flaws even in software that has already received extensive scrutiny.
Cost was also a factor. Static analysis across the 10 stacks cost about USD $935, or an average of USD $2.58 per finding.
According to the research, binary analysis was more expensive and slower than source-code analysis, costing roughly six times as much and taking twice as long. Even so, the findings suggest AI-assisted vulnerability research could be applied to proprietary firmware where source code is not available.
Critical systems
The findings matter because TCP/IP stacks sit deep inside a wide range of connected products and can be reused by multiple manufacturers. A weakness in one stack can therefore spread through the technology supply chain and affect many downstream products.
This creates a difficult problem for critical infrastructure operators, especially when devices remain in use long after their software was written. In many cases, a fix must pass from the original software developer through firmware and product vendors before it reaches the organisation using the affected device.
For unsupported or discontinued products, a patch may never arrive. That leaves operators relying on other measures, such as network segmentation and device isolation, to limit exposure.
"AI did not eliminate the need for expert vulnerability researchers, but it dramatically compressed the work," said Daniel dos Santos, Vice President 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," dos Santos said.
Operational response
Forescout said organisations should prioritise risk based on reachability, exploitability, and operational impact rather than treating every vulnerability the same way. It also advised reducing unnecessary exposure by disabling services such as FTP, Telnet, IPv6, and web administration interfaces where possible.
It also recommended isolating end-of-life or unsupported devices in tightly controlled network segments when patching is not possible. The research further urged faster patching cycles for embedded devices where fixes do exist, reflecting a faster pace of vulnerability disclosure driven by AI.
Procurement was another area highlighted in the research. Buyers should require suppliers to define support lifetimes, publish vulnerability disclosure policies, provide software bills of materials, and show evidence of ongoing security testing.
Barry Mainz, Chief Executive Officer of Forescout, said the practical challenge for many operators remained unchanged despite the rise of AI-assisted discovery.
"The speed of vulnerability discovery is increasing with AI, but the reality for asset owners and operators hasn't changed," Mainz said.
"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," Mainz said.