Bias in Hiring is as Old as Hiring Itself.
New Technology Might Change That.
As the story goes, the modern job interview process goes back to 1921 and Thomas A. Edison. Yep, that Thomas A. Edison. He developed a questionnaire with about 140 questions about general knowledge topics and administered it to all candidates for positions at any of his companies. Questions included “Who composed ‘Il Trovatore,’” and “Where are condors to be found?” Edison apparently graded the questionnaire himself, and only about 2% of candidates passed. Even Albert Einstein (yep, that Albert Einstein) failed the test at one point because he couldn’t remember what the speed of sound was.
Edison’s test was of his own making, and it was likely biased towards hiring people just like him. He reasoned that anyone who knew the same sorts of trivia he knew would be good at memorizing information, and that was, in turn, good for making quick business decisions. Interview experts today would call this “affinity bias,” and it’s one of the most common biases found in the hiring process.
What’s the problem with bias in hiring?
In hiring, conscious bias (overt, deliberate bias against a subset of the population) can get you into legal trouble pretty quickly. The EEOC governs what you can and can’t ask in interviews when it comes to protected traits like gender and disability status. Unconscious bias (stereotypes or assumptions about certain populations) gets you into less legal trouble but still gets in the way of hiring diverse, high-performing teams. Biased hiring means you’ll frequently get a homogenous team that lacks innovation.
Many tools have been developed to address this bias in the past few years. Applicant tracking systems can now anonymize résumés, and scanners have been developed to detect gender bias in job advertisements. Skills-based hiring uses automated testing to provide objective scores of hard abilities that are highly predictive of future job success—when implemented correctly.
The Advent of Generative AI.
With ChatGPT revolutionizing how we work with words, there are new possibilities opening up to reduce bias even further. AI can quickly and accurately review résumés to find potential matches. It can be used to create automated skills tests to encourage more skills-based hiring, and it can automate the task of building interview guides—long documents filled with interview questions and rubrics—that help to make the interview process more objective.
So far, though, the track record for AI in hiring isn’t great. Technologies like AI text and video interviewing are wildly inaccurate in recent tests. Amazon infamously developed an AI résumé reviewing system that consistently downvoted women because of the way their model was trained. Other résumé reviewing systems introduced issues because of selection bias—hiring managers were told to find the high performers on the team and almost invariably chose men.
A Brighter Future.
But it’s not all doom and gloom. At Gordian Knot, we’re working on new algorithms that guard for bias from the outset and ensure that résumés are fairly assessed. We use contextual mapping to get insight into what skills a candidate might possess even if they don’t specifically list it on their CV. Companies like HireGuide are developing tools to create bias-free interview guides and AI coaches to help interviewers ask better and more relevant questions.
By shifting work onto a generative AI, bias can be ameliorated considerably, but only with careful, deliberate work. GPTs are trained with a wide array of data, so care must be taken in the instructions provided to them when attempting to use them to solve problems in the hiring space.
Conclusion
Bias in hiring may be as old as hiring itself, but we’re entering a new era where technology — if carefully designed and implemented — can help level the playing field. Generative AI, while not without its challenges, has the potential to make hiring more objective, reducing unconscious bias and focusing on skills and cultural fit over surface-level assumptions. However, the key lies in how we build and train these systems, ensuring that they reflect the diversity we strive for in our teams, rather than reinforcing old biases.
At Gordian Knot, we’re committed to shaping a future where technology works for us, not against us. By leveraging AI to fairly assess candidates and create more inclusive hiring processes, we believe we can help companies build stronger, more innovative teams. The road ahead requires thoughtful innovation, but with the right tools and careful attention to bias, we can move toward a hiring process that’s as inclusive as it is effective.
The future of hiring is not about eliminating bias entirely — it’s about recognizing it, addressing it, and designing systems that give everyone a fair shot. With the right approach, AI can help us get there.
