Having recently run an internal training session on unconscious bias, I took time to ponder the future implications of AI in Talent Acquisition with assisting in eliminating the negative impact of bias in the recruitment and selection process.
What is unconscious bias?
Bias is a person’s inclination or prejudice against another person or group of people. Unconscious biases are the prejudices every human has and acts on without thinking or malicious intent. This may cloud or blind our logical judgement when making decisions during the recruitment and selection process.
So, is Artificial Intelligence the remedy to overcoming unconscious bias and promoting inclusivity?
Taking a simplified view, the argument for AI in TA would seem an easy argument to make – you simply isolate human decision-making processes that are open to the risk of a human exercising unconscious bias and replace with AI to eliminate the bias. The key decision-making points during a recruitment process begin with the creation of the job advert and job description. An example of AI at this stage can be demonstrated with Textio, a text editor that can be used to rewrite job descriptions to appeal to under-represented groups. This had reportedly contributed to a 47% increase in female applications for software giant Atlassian.
A further application of AI during TA may exist at shortlisting stage, with a number of products available that combine human and artificial intelligence to shortlist job applicants based on previous preferences.
Here’s the rhetoric – Does this take the “Human” out of Human Resources?
Perhaps, although a recent social media post seems to have capture the fallible nature of AI. In August 2017 Chukwuemeka Afigbo had posted a video to Facebook demonstrating an interesting flaw of a seemingly simple and domesticated application of AI – A soap dispenser!
Chukwuemeka Afigbo, Facebook’s Head of Platform Partnerships posted a video of a Caucasian male placing his hand under a soap dispenser and soap being release on the first attempt. However, when a darker-skinned male does the same, no soap is released. A white paper towel is then placed under the dispenser, which works immediately.
So are the Algorithms just as good as their creators?
What was learned from the short clip is that AI isn’t the issue – but the major lesson learned is that important further steps are required in diversifying the demographic of the professionals and collaborators involved in the creation of AI and acknowledging that the majority of accessible data sets may not be representative of the end user.
Using AI in TA doesn’t end with creating diverse teams, design teams should be cognisant of functionality being inclusive. Its safe to say that AI is continuously developing, and TA remains an everchanging discipline within HR – the harmonisation of AI with human TA processes continue to change and shape hiring processes across the industry.
For more more information email arun.bhatia@methods.co.uk.