AI Pentesting Overload: Teams Can't Validate Findings

AI-assisted penetration testing tools are creating a backlog for security teams. New research shows these tools generate vulnerability findings faster than teams can verify them. A survey of 158 security practitioners found that 87.8% said AI output required significant manual validation. Over a quarter of AI-generated findings needed rework before being trusted. Many respondents described tools returning hundreds of findings that were duplicates, false positives, or even fabricated. One practitioner noted an AI tool produced 300 findings, with 250 later found to be "junk." The leading frustration was fabricated or hallucinated findings. This leads to a "trust effect," where teams become more cautious, increasing verification work even for genuine findings. This means AI, intended to save time, is often creating more manual work.