Jiransecurity Jaehyung Park Strategy & Planning Manager

AI- and ML-based threat detection is advancing rapidly. However sophisticated it grows, though, flawless preemptive defense stays out of reach: the judgment and the action of the final moment still rest with people. So how should we manage the security threats that arrive through people, namely human risk? This session proposes treating human risk not as a verdict on individuals but as a measurable degree of threat, assessed transparently and in fine-grained detail against open standards such as MITRE ATT&CK and the NIST Phish Scale, and then managed through a continuous cycle of measure, improve, and measure again. Paradoxically, the more security becomes autonomous under AI agents, the more it matters who sets the standard for what good security looks like, and that remains a human task.