The Most Qualified Candidate Doesn't Always Get the Job with Trent Cotton
カートのアイテムが多すぎます
カートに追加できませんでした。
ウィッシュリストに追加できませんでした。
ほしい物リストの削除に失敗しました。
ポッドキャストのフォローに失敗しました
ポッドキャストのフォロー解除に失敗しました
-
ナレーター:
-
著者:
Why more experience doesn't always mean more skills, and how artificial intelligence can help us challenge hiring biases.
What happens when the candidate who demonstrates the strongest skills doesn't get the job?
In a new episode of The New Mainstream Podcast, ThinkNow Co-Founder Mario Carrasco speaks with Trent Cotton, Head of Talent Insights and Analyst Relations at iCIMS, about how data and artificial intelligence are transforming recruitment and, more importantly, how they can help us identifybiases that have long been considered a normal part of the hiring process.
When Experience Outweighs Skills
During the conversation, Cotton shares an experience from his career that illustrates how bias can influence hiring decisions even when organizations have objective information about candidates' abilities.
At a previous organization, he implemented an AI-powered technical assessment to evaluate how candidates approached problems related to software development and technology.
For one particular position, the team was looking for a professional with approximately seven or eight years of experience.
Two candidates stood out during the process.
The first had around eight years of experience and scored 68 points on the technical assessment.
The second had approximately two years of experience, had completed a two-year technical education program, and had developed much of his expertise independently through resources such as GitHub, YouTube, and specialized communities. He scored 96 points.
Initially, the team favored the candidate with the higher score. However, after discovering that he had less professional experience, they decided to hire the candidate with eight years of experience.
Three months later, the company had to remove him from the project because he was unable to perform the job effectively.
Hiring Bias Existed Long Before AI
Much of today's conversation about artificial intelligence in human resources focuses on the risks of algorithmic bias. It's a legitimate concern, particularly when systems are trained on historical data that may reflect existing inequalities.
However, Cotton invites us to consider another perspective: traditional hiring processes are not free from bias either.
One example he highlights is recency bias.
Imagine a company posts a job opening and receives 350 applications in less than a day. Although every candidate could potentially be evaluated, the volume of applications often leads recruiters to review only a portion of them.
According to Cotton, a recruiter might review between 20 and 50 applications before selecting the candidates who will move forward.
But what happens if the most qualified person applied shortly after the position was posted and wasn't included in that group? Their application might never be considered, not because they lack the necessary skills, but because of how thereview process was organized.
Something similar happens during interviews. An outgoing candidate who communicates confidently and builds an immediate personal connection may leave a strongerimpression than someone who is more reserved, even when the latter has greater technical expertise.
In both situations, human perception can influence decisions that should be centered on the abilities required to perform the job.
Artificial intelligence does not eliminate bias on its own. Its value depends on how it is designed, implemented, and supervised.
The problem was that expectations aboutexperience ultimately outweighed the evidence of actual skills.