Can AI really assess human competencies?

Discussions about using AI to assess people often begin with the wrong question: “Can AI make the right decision about a person?”

I think a better question is: Can AI help us gather better evidence of human behaviour and interpret that evidence more consistently?

This distinction underpins the systems we are working on at The Leapers.

Instead of asking an algorithm to label someone a “good leader”, “analytical” or “high potential”, we first define the competency we want to assess. We then identify the observable behaviours that demonstrate it and place people in realistic scenarios where they can show those behaviours.

AI is a tool for processing the behavioural evidence that emerges from these situations.

It can help structure written and spoken responses, reveal recurring patterns across exercises, and connect observations with predefined behavioural indicators.

The final interpretation remains with people.

I believe AI's strongest contribution to human assessment is to help us see more clearly.