Skip page header and navigation

Addressing AI bias in recruitment technology

Staffing Stream

Addressing AI bias in recruitment technology

Audra L. Woods
| October 21, 2024
Image
AI ethics or AI Law concept

main content

For decades, we’ve looked to workforce technology to advance how we source, engage and manage talent. AI, and particularly generative AI, is the latest technological advancement to shift how the global workforce gets work done. It is surfacing across several applications that recruiters use daily — proving to save time and increase overall productivity in the recruitment process. AI is being used to screen résumés, conduct initial job screenings and video interviews, and support various onboarding tasks. However, with these tools, candidates could advance through the hiring process without touching that crucial human element. AI can result in significant acceleration of processes but can also be of concern when considering various biases that could be built into the software.

In my role as a diversity, equity and inclusion marketplace leader, I’ve worked with dozens of industry leaders and staffing firms who have expressed caution when it comes to relying too heavily on AI due to reported hallucinations and bias.   

Defining AI Bias

AI is created by humans, and it is a basic truth that humans have bias. This has inadvertently affected the tools that we use to review, analyze and select the candidates that apply for work.

An IBM article defines AI bias as “AI systems that produce biased results that reflect and perpetuate human biases within a society, including historical and current social inequality.” With the advances we are experiencing in AI, it is feasible that a candidate may not connect with another human on their team until the second or third interview — and in some cases even later. An unknown bias could eliminate candidates that may have been ideal. If the developer of the tools has inadvertently embedded coding that overweigh certain factors in the algorithmic decision-making, or the organization inputs historical data that has underlying systemic bias against equity-seeking populations, great workers could fall through the cracks.

In addition, uploading historical data of previous hires could further integrate bias, as the machine learning will base its decision-making on what has come before. This can impact diversity as well as salary disparities. 

With the growing need for AI adoption to stay competitive in the market, it is imperative that companies learn ways to mitigate biases and continue to improve machine learning.

Ways to Mitigate Bias Risks in AI

It is important to protect the organization from the repercussions of biased hiring practices and the concerns that could arise with a lack of human interaction during the recruiting, interviewing and onboarding processes. In addition to finding opportunities for human touch points, there are other steps that businesses can take to mitigate bias.

Research the technology and company. Has the company had a history of lawsuits, discrimination complaints or non-equitable practices? Have other companies experienced biased results from using that software or platform?

Analyze your internal data. Review your company’s hiring practices, identifying historical trends and any potential barriers or disparate impacts.

Conduct frequent audits. Once you’ve vetted the technology and your data, you will need to conduct regular audits to identify any potential bias in the AI process and which candidates are not getting selected to identify any trends.

As well as mitigating bias risk, adding human checkpoints to your process could put candidates at ease, making them feel your organization sees them as an important individual who can make an impact and not just a number. These regular human interventions also ensure that each section of your recruiting process is transparent to your leaders and aligns with your organizational values and goals.

Making AI Work for You

AI is an integral part of the future of work, and it is crucial that you can find ways to make it perform for your business. With small adjustments to how you source talent, you can benefit from the increased productivity, speed and computing power of AI while still ensuring you are hiring the best candidates to empower your workforce.