AI & Talent Technology
Interestingly we still seem to follow historic patterns within talent technology and AI, believing that the “next new thing” will solve all the issues of recruiting talent for your business. People are complex things and our feverish drive to automate and abdicate decision making is losing sight of a core element – we are dealing with people and people have bias.
Sure we can automate certain steps within the hiring process and present better data to make quality hiring decisions but that’s the kicker – we still need to make decisions, not the system.
The hype around AI is a front of mind example and the Amazon experience on how bias can be automated is just one example of the cracks appearing on a number of fronts.
Lets quickly get some hype out of the way by just having a quick view of where AI fits in
Evolution of AI
Around since the 50’s in various forms its bandied around as the catch-all phrase that all new products apparently must have. In recent pitches I attended in London I was told “just say you have an AI-driven piece of technology even if you don’t ” …brilliant!
So how does this relate to bias? Well as increasing examples show we are simply inserting bias data sets into our shining new “AI-powered” tools and wondering why the very imbalances we are trying to right are in fact growing.
With over 180 Human biases in play, we can’t be expected to mitigate all.
Cognitive Bias Codex
What Can Be Done?
However what we can do is take the blinders off and, if committed to removing bias, start by simply ensuring that you acknowledge that its there in your business (all businesses in fact).
Recruitment teams are a powerful channel for changing bias and unfortunately, they have also been the greatest perpetrators (not us! I hear already). Part blame can be the metrics they are measured on being focused on processing speed, and the larger part being fear embedding these behaviors.
· Fear of losing a client or offending a manager if you say “sorry I can’t recruit for a client/manager that only wants white/male graduates from a certain socio-economic group you belong to”
· Fear of not hitting their numbers so when faced with long applicant lists – they go with the ‘safe’ ones
· Fear of pushing back when they know something “just isn’t right”
So, step two, ensure you have a culture that supports and encourages robust discussion around bias in a no blame way.
I believe that the sector focus needs to be more on rehumanising the recruitment journey and we have been distracted somewhat by promises that are already ringing hollow.
Ensuring that your processes don’t support bias and your team has the training and support to counter it should be mission critical and your greatest investment as this is where real change can happen.
Only with these foundations in place can you confidently explore technologies to support your team in removing bias.
So when assessing technologies perhaps a couple of questions to throw in the mix
· Do we have the foundations in place to ensure we are enhancing our capabilities with this technology? Or are we simply putting diesel in an electric car and wondering what’s wrong.
· Are we replacing decisions that should remain in the human domain with automation? – Why are we doing this? And how will that impact?
We are dealing with people, people have bias – If we put more focus on removing the bias people have, the machines will follow our algorithm ….we hope!