Human Context Remains Key to AI-Assisted Vulnerability Research

Summary

The discussion cites a survey rating creative manual research that finds issues missed by automation at 6.3 out of 10. It argues that application context helps researchers and AI agents assess which vulnerabilities matter most.

Key points

• In programs using AI tools, survey respondents rated the share of accepted findings from creative manual research missed by automation at 6.3 out of 10 on average.
• The speaker says the gap is shrinking, but application context helps determine whether a vulnerability or misconfiguration matters to a specific program.
• The discussion says AI agents need program and application context, including what would cause the most harm if exploited.