SecurityBrief US - Technology news for CISOs & cybersecurity decision-makers
United States
Forescout AI finds 99 flaws in networking software

Forescout AI finds 99 flaws in networking software

Fri, 20th Nov 2026 (Today)
Mara Sugue
MARA SUGUE News Editor

Forescout has published research showing that artificial intelligence helped its researchers find 99 vulnerabilities in networking software, reducing analysis time from about nine months to 90 hours.

The findings came from tests on 10 TCP/IP stacks, software components that connected devices use to communicate over networks. These stacks are widely embedded in operational technology, internet of things devices, medical equipment, network gear and other cyber-physical systems.

Of the 99 validated vulnerabilities, 66 were previously unknown and 33 were already known but rediscovered by the AI-assisted system. More than half of the new vulnerabilities, 52%, were rated high or critical severity, and 54 were reported to vendors and maintainers.

Forescout said the automated static analysis cost USD $935 across the 10 stacks, which it calculated as an average of USD $2.58 per finding. It added that binary analysis was about six times more expensive and took twice as long as source-code analysis, though it still showed the same approach could be applied to proprietary firmware where source code is unavailable.

The research examined whether generally available AI models could help identify vulnerabilities in both open-source software and proprietary binaries on a limited budget. Researchers built what Forescout described as a specialised AI-assisted vulnerability research harness and applied it across seven stacks already studied by its Vedere Labs unit, as well as three additional stacks.

Infrastructure risk

The results matter because the same TCP/IP software components often sit inside products sold by multiple manufacturers. That means a single flaw can affect many downstream products across the technology supply chain, particularly in environments where devices remain in service long after vendor support ends.

Those conditions are common in critical infrastructure, where patching can be slow and equipment may stay in operation for decades. Even when developers issue a fix, updates often need to pass through firmware suppliers and product vendors before reaching the organisation operating the affected device.

Daniel dos Santos, Vice President of Research at Forescout, said the work showed how AI changes the pace and cost of vulnerability research without replacing specialist analysts. "AI did not eliminate the need for expert vulnerability researchers, but it dramatically compressed the work. 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, prioritize, 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.

Forescout also said the AI-assisted system rediscovered 40% of previously known vulnerabilities in software that had already undergone extensive manual research. It argued that this suggests AI tools can still find flaws in targets that have been examined before.

Operational response

Barry Mainz, Chief Executive Officer of Forescout, said faster vulnerability discovery does not change the practical constraints facing operators of legacy systems. "The speed of vulnerability discovery is increasing with AI, but the reality for asset owners and operators hasn't changed. 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 organizations 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.

Forescout advised organisations to assess vulnerabilities according to reachability, exploitability and operational impact rather than treating all flaws the same. It also recommended reducing unnecessary network exposure, isolating unsupported devices that cannot be patched, shortening patching timelines where fixes are available, and requiring suppliers to provide support lifetimes, software bills of materials and vulnerability disclosure policies.

The research adds to a broader cybersecurity debate over whether off-the-shelf AI will lower the barrier to entry for both defenders and attackers. Forescout said its findings indicate that vulnerability discovery in embedded and foundational software can now be carried out much more quickly and at relatively low cost, including in software used widely across critical infrastructure sectors.