Diversity is a very prevalent and relevant issue in our society today. It is even more important to consider in the business environment. Business has historically been an exclusive sector of the workforce and because of this, we are still seeing the repercussions today.
Today 48% of businesses are not on track to meet their diversity goals, this is not an acceptable statistic. The choice to embrace diversity should not have to be quantified but studies show that the 20 most diverse companies in the S&P 500 have higher average profitability than their less diverse counterparts.
Embracing diversity must come from an executive level and many executives recognize and support this. 58% say it is a social corporate responsibility, 53% admit that it contributes a wide variety of skill sets, 43% say that diversity increases customer relationships.
Not only is diversity a social responsibility, but it also improves business outcomes. Companies that are gender diverse outperform on profitability and value creation versus companies that have a majority gender. Racially diverse companies set an industry level of profitability and excel in financial performance.
How AI Contributes to the Problem
However, the lack of diversity may not be due entirely to human bias. Many recruiting firms use AI or artificial intelligence to pre-screen candidates to help reduce the amount of time a recruiter has to spend looking at applications. As coded entities, AI software is only as good as the person who programmed it.
This can lead to several AI-based bias cases. One of the most common is Boolean search bias. This occurs when the AI relies on various programmed keywords to find a result. These keywords can perpetuate bias through the person who programmed them.
AI bias can also be a result of recruiter skill. Creating strong search queries requires a great deal of skill. AI software that is programmed by a newer person may lack some of the qualifying terms that a professional may use.
Another flaw with AI recruiting is that individuals may write their resume based on the AI specification. If a candidate knows what the machine is looking for, they may write a resume that does not actively reflect their skills just so they can get the job.
One of the final biases that can be committed by AI software is synonym searching. AI machines may not be trained to search for synonyms to exclude individuals with similar qualifications but not the exact ones.
Why AI Isn’t the One to Blame for the Lack of Diversity
Many of the AI biases stem from human ones. One human bias is called the Halo and Horns effect. This is an instant bias or assumption that we make of a person based on one negative action. We then take this negative action and generalize all the actions of that person.
Human bias can also occur in the contrast effect. If a candidate messes up an interview or does not meet the requirements, we might compare the next candidate to that one. Even if the candidate is only nominally better than the other candidate we associated them with the idea of being the best.
One of the final human biases is called the central tendency effect. This occurs when we side with the general group opinion rather than stating our open. For example, when interviewing a candidate you may have severe reservations but if every other person in the office likes them you will choose to bury your reservations and side with group opinion.
While AI may be the solution to helping eliminate bias in the recruiting process and the workplace there is still a human element to it. To create a bias-free work environment, we must be willing to face and accept our natural biases.
Not only must this be done, but we also have to actively come up with solutions to help eliminate the occurrence of bias and the lack of diversity within the workplace. To be effective, we must include all backgrounds in this conversation instead of excluding some. include all backgrounds in this conversation instead of excluding some.