Is AI making our recruitment fairer or not? We need to keep talking about it

Will A.I. improve or worsen fairness in recruitment? Now that almost 90% of companies say they use some form of A.I. in their recruitment and selection, this question is becoming increasingly important. Whilst some believe that algorithms will make the recruitment process fairer by reducing human bias and ‘noise‘ in decision-making, others warn that algorithms can reproduce and even reinforce existing inequalities at scale. However, according to a number of (Dutch) researchers, both camps ignore a crucial reality: A.I. implementation itself changes the definition of what is fair.

‘The risk of A.I. in recruitment is quiet, but serious.’

And that is a problem, they write in Harvard Business Review. Most companies implement A.I. recruitment tools without thoroughly and continuously evaluating them during development and use. ‘Fairness is seen as something the tool delivers from day one, rather than something leaders need to monitor and adjust in practice. The risk is quiet, but serious: A.I. can narrow the candidate pool, sideline human expertise and lock in one definition of fairness’, according to Elmira van den Broek (of Stockholm University, pictured), Marleen Huysman and Anastasia Sergeeva (both of VU Amsterdam).