IE 11 Not Supported

For optimal browsing, we recommend Chrome, Firefox or Safari browsers.

AI in Hiring: Can It Reduce Bias?

As governments increasingly fold artificial intelligence into their hiring processes, the question emerges: Does AI increase or reduce bias? Some experts argue that, when implemented responsibly, it reduces bias and directs a focus on skills.

Closeup of two people shaking hands with the outline of a human head with "AI" in the brain superimposed over the image.
There are risks in the use of artificial intelligence (AI) in hiring, but some experts argue AI technologies can actually reduce biases in the process.

AI has powered hiring behind the scenes for years; and while some say this use case can be a harmful one, others make the case that it can improve the process — potentially eliminating bias and directing a focus toward candidates’ skills.

Career Ready DC, an AI-powered career platform launched in December by the Washington, D.C. Department of Employment Services (DOES), aims to do exactly that. In announcing the platform, D.C. Mayor Muriel Bowser emphasized the technology could help make the job search process more equitable.

The platform uses algorithms to highlight job matches, removing language that could perpetuate bias and instead promoting a person’s skill set. The goal is to level the playing field for all people to equitably compete for job openings.

“For so long, in marginalized communities, technology hasn’t really been used to innovate or advance services,” DOES Director Unique Morris-Hughes told Government Technology. “And so, what Career Ready DC provides is another tool to eradicate bias.”

The platform’s launch came through a partnership between DOES and Eightfold AI, which helps organizations address workforce challenges with AI. Dan Hopkins, the company’s vice president of global public sector and applied AI, told GT the use of AI in something as potentially sensitive as hiring is a “critically important topic,” noting that scrutiny of use in this area is a good thing as it helps ensure AI will be used as “a source of good.”

What sets Eightfold AI apart, Hopkins said, is that it helps organizations implement a skills-based approach to hiring. The approach uses AI to understand skills revealed both by a job seeker’s previous roles and education, as well as potential skills based on job trajectory information. The platform then matches those candidates to employers.

Hopkins said the tool allows the humans that are part of the hiring process to leverage profile masking — seeing skills rather than factors like name, gender and race — to focus on skill-related data and obfuscate data points that could lead to biased decision-making.

“So, you’re still having humans make those decisions, but you’re giving them a more informed data set to make the very best decision,” he said.

This, he said, will improve the process for job seekers as well, who often interpret skills requirements through vague job descriptions. Hopkins cited a Harvard study that found women don’t apply to jobs unless they meet all qualifications, which means diverse and qualified candidates often filter themselves out.

Organizations using Eightfold AI software can access dashboards and reports to understand trends — or even drop-off points — during the hiring process, he said. For example, if female candidates are dropping off after on-site interviews, the organization can identify that trend and take action to course correct.

DOES is exploring other tech-related, resident-facing tools to further support job seekers in the process, including a forthcoming training provider scorecard so those looking to upskill can get data about training providers.

And across the nation, the trend of skills-based hiring is gaining ground, with many government organizations expanding hiring channels. Hopkins referenced an executive order on hiring under the Trump administration; and an executive order on diversity, equity, inclusion and accessibility under the Biden administration as two actions largely driving this shift: “And this is where artificial intelligence can really be a game-changer.”

Eightfold AI is not the only tool being used in this space. HireVue is another technology company aiming to improve the hiring process for organizations.

The company recently released a report revealing that 66 percent of the employers surveyed are using AI-powered skills assessments as part of the hiring process. With or without AI, the report found 96 percent of respondents are increasing their focus on skills.

Juan Salazar, VP of public sector at HireVue, told GT via email that the mitigation of human bias in hiring is one of the primary benefits of using AI.

Salazar said HireVue helps public-sector organizations evaluate candidates based on skills, potential and interests, rather than unnecessary educational and credential requirements. With HireVue’s technology, name, gender, age and ethnicity are never used to assess candidates.

“HireVue customers often see improvements in diversity even when they’re not prioritizing that as an outcome,” he wrote.

But Salazar also said there is a need for greater regulation in the area of AI in hiring — something he said the company supports. While he argued that AI can have benefits in the hiring process, he specified the importance of using "appropriately trained and audited algorithms."

Based on its potential, Hopkins noted AI is gaining popularity in this space; but he, too, urged caution in its development.

“When you … train your AI off of your data set, any bias that is part of your organization is going to be re-perpetuated in the outcomes of the new AI tool,” he said, noting the value of integrating a more expansive, external data set in addition to that of an organization.

Morris-Hughes, the DOES director, said department case managers will monitor Career Ready DC, in ongoing collaboration with a DOES team of IT professionals, to ensure data integrity and that the platform is being used as intended. She also cited Bowser's February actions to ensure the responsible use of AI in the District of Columbia, explaining that the mayor’s order outlines both procedures and values for responsible — and equitable — AI implementation.

However, the platform’s launch took place prior to Bowser’s order. As such, Morris-Hughes underlined the importance of pre-work that the DOES team did to ensure the tools the department was investing in allowed transparency, security and equity. Morris-Hughes said the department had gained essential experience investing in new tech tools from its role running the unemployment insurance system during the COVID-19 pandemic.

According to Morris-Hughes, support from city leaders provided the crucial foundation for the platform’s creation, and she credited the district’s centering of equity in its vision and values. She urged other government organizations that are exploring similar AI uses to work with their CIO to do so.

“I’m not afraid; I’m encouraged by the use of technology and AI, and how we can use technology for marginalized communities — to eradicate poverty, to level the playing field, to get people family-sustaining jobs,” Morris-Hughes said. “It’s actually something to be quite excited about.”
Julia Edinger is a staff writer for Government Technology. She has a bachelor's degree in English from the University of Toledo and has since worked in publishing and media. She's currently located in Southern California.