Why Human Judgment Matters More As AI Gets Better, Faster

Haris Pylarinos is the founder and CEO of Hack The Box, a leading cyber readiness platform for the AI era.

The better and faster AI becomes, the more dangerous it is to assume we need less human expertise.

That may sound counterintuitive. AI can already analyze enormous volumes of data, surface anomalies, summarize technical issues and recommend actions faster than any human team. As those capabilities improve, it is tempting to conclude that experience will matter less, but I believe the opposite.

AI is getting remarkably good at identifying problems and recommending responses. What it cannot reliably judge are the operational consequences of a decision within the context of a particular business. As we hand more routine work to machines, people will increasingly be left with the difficult, ambiguous decisions that data alone cannot resolve.

In cybersecurity, the technically obvious fix isn’t always the right business decision. Knowing the difference requires hands-on experience.

The Value Of Knowing The Context

Earlier in his career, a colleague of mine was a penetration tester assessing a telecommunications billing environment. As a security control, access to an FTP service was restricted to a single concurrent session because only the designated billing system was supposed to connect and retrieve the data. During his first assessment, the entire team was not aware of this dependency. As part of the audit, he connected to the FTP service and established a session. Because the service allowed only one active connection, the legitimate billing process could no longer connect and retrieve its data.

In today’s AI universe, AI understands the vulnerability. The experienced professional understands the trade-offs and potential consequences of each response.

​This is why simply putting a human “in the loop” is not enough. If that person lacks the judgment to challenge what the machine is telling them, oversight can easily become little more than approval.