When the Artefact Stops Being the Evidence

AI is creating an interesting contradiction in recruitment.

We increasingly expect people to use AI at work, while some assessment processes are still trying to work out how to stop candidates using it before they get there.

I suspect that becomes harder to sustain.

CVs, written applications, presentations and take-home exercises have always been proxies. We use the finished artefact to infer something about the person who produced it.

AI weakens that inference.

A very polished answer no longer tells us quite as much about how somebody arrived at it. But banning the tool does not necessarily solve the problem if using the tool well will be part of the job.

Perhaps the assessment needs to move closer to the work itself.

Give someone a difficult problem. Let them use AI. Then understand how they work.

What did they ask it to do? What context did they give it? What did they decide not to delegate? Which answer did they challenge? What did they reject? Where did they bring their own expertise to bear? Did they recognise uncertainty? Could they explain the final decision and stand behind it?

And perhaps most importantly: when did they decide not to use AI?

That feels much closer to the capability many organisations will actually need.

Interestingly, it could make recruitment more human rather than less.

If producing the artefact becomes easier, the conversation around it becomes more valuable. We need to understand the person’s reasoning, judgement, curiosity and ability to challenge what a machine gives them.

There will still be assessments where unaided capability is exactly what we need to measure. The point is not that every test should allow every tool.

The point is that we should be clear what the test is for.

If the role expects somebody to work with AI, then assessing how well they work with AI may tell us more than pretending the technology disappears for the duration of the interview.

The finished artefact is becoming a weaker signal.

The way someone works may become a stronger one.