Are we really putting our faith in Artificial Intelligence? Is it really that good yet? I don’t think so.
I have been involved in recruitment for well over 30 years. I have seen plenty of successful and unsuccessful recruitment strategies during that time. But I do have to question the legality, the effectiveness and indeed the morality behind the current vogue for using artificial intelligence in recruitment.
Please do not get me wrong. I am an avid supporter of technology, we had our first website in 1995 only three years after they had invented the world wide web. And it transformed our business.
And I understand the need for organisations to find a way to reliably filter applicants from a torrent of applications in the current economic climate.
But the way that some of these systems are being applied is truly shocking. I heard the other day that a reputable, highly respected multinational, was using artificial intelligence in automated interviews for senior management appointments.
And just like in Little Britain, when “Computer says no” just two minutes after the interview, you have to question the result. Especially as the individual was actually known to the organisation already, and they interviewed them anyway.
New technology is great, but just like vaccines have to go through an extremely rigorous testing process, so does something as fundamental as a recruitment system.
Just because you don’t have enough positive, action words in your CV does not mean you are a bad candidate. Just because the facial recognition software didn’t see you smile enough, accentuate enough words, use the right vocabulary at the right time, does not mean you do not know your job.
And what about feedback loops? Using algorithms to determine from those selected who performed best during the interview process and even who finally got selected.
If the recruiters themselves are discriminatory afterwards, then that feedback loop will automatically go back in and be discriminatory at the sifting stage.
There has been much talk about unconscious bias. What about AI bias based on unconscious bias? Surely that’s even worse as it is hidden, cannot be discovered, and because no individual has been responsible, it is difficult to litigate against.
So we do need to think carefully. Plenty of suppliers will say they have this covered, but how can you tell? Recruitment is a very human activity, just like dating. And I read about plenty of complaints about dating apps – and they have been doing this a lot longer.