AI Testing Tools: Cybersecurity Pros Losing Trust in AI

Cybersecurity professionals are losing trust in AI testing tools to find vulnerabilities. Less than one in ten now trust these tools. Over three-quarters of these professionals report that their AI vulnerability scanning tools missed critical flaws. Confidence in fully automated AI testing has dropped significantly, from 29% in 2025 to just 9% this year. As a result, a hybrid model, combining automation with human expertise, has surged to 47% adoption. Experts emphasize that automation should complement, not replace, human skills in uncovering complex risks. This shift is happening because automated tools are missing critical vulnerabilities, and the mean time to resolve AI security issues has nearly doubled, from 19 to 36 days. This matters because it highlights the ongoing need for human oversight in critical cybersecurity functions, even with advancing AI.