AI-Related Software Vulnerabilities: 2025-2026
Barracuda Networks, Wednesday, August 12th, 2026
Barracuda research finds AI-related CVEs are more severe and more likely to be exploited than other flaws.
Barracuda published research on AI-related software vulnerabilities across 2025 and 2026, finding that AI-related vulnerabilities are generally more severe and more likely to be exploited than those affecting other software.
The report notes that AI-related CVEs are growing in number. The findings underline the security risks accompanying rapid AI adoption. Barracuda pairs the analysis with practical guidance for developers and organisations seeking to secure AI-powered applications and infrastructure.
The work is aimed at teams whose attack surface is expanding as AI components are added to existing stacks.