AI Integration Reduces Vulnerability Discovery Time to 90 Hours

Research indicates that the application of artificial intelligence to vulnerability discovery has significantly compressed the time required to identify security flaws in critical infrastructure. By utilizing an AI-assisted research harness, analysis time for complex TCP/IP stacks was reduced from approximately nine months to 90 hours. This process identified 99 vulnerabilities, 66 of which were previously unknown, with over half classified as high or critical severity.

The automated analysis was conducted at a total cost of approximately $935, averaging roughly $2.58 per identified vulnerability. This shift in the economics of cyber research enables both security teams and malicious actors to identify and exploit weaknesses in embedded software at an accelerated pace. Because TCP/IP stacks are foundational to operational technology, IoT, and medical devices, a single vulnerability can propagate risk across extensive supply chains.

Operational risks remain elevated due to the prevalence of legacy systems that are difficult to patch or remove from service. Experts suggest that organizations must prioritize risk containment strategies over traditional patching cycles to address the increased velocity of threat discovery. The findings underscore a broader trend where AI is compressing the reconnaissance phase of cyber attacks, particularly against strategically significant targets.

Geopolitical threat activity targeting the United Kingdom has also intensified, with 108 distinct threat actors identified in 2026. Hacktivist activity has seen a 337% increase compared to 2025, with a significant portion of these claims linked to pro-Russian groups. State-sponsored actors and cyber criminals continue to focus on government institutions and public services, leveraging the increased speed of vulnerability identification to target critical infrastructure.