The technology we create is a reflection of ourselves. That includes our own systemic prejudice. How can we ensure that AI at work is a force for good?

Left unchecked, AI has the potential to build a future workforce that is less inclusive, less diverse and that sustains barriers to entry for disadvantaged groups. But, when used responsibly, AI can shape the future of work for the better. It can counteract the very biases and barriers that job seekers face today.

Speaking to an audience of hundreds of HR professionals at Indeed FutureWorks 2023 on September 21, Indeed SVP of Environmental, Social and Governance LaFawn Davis tackled the weighty topic of bias in AI, as well as how Indeed’s Responsible AI Team is committed to making work better and more equitable for all.

Professional headshot of LaFawn Davis, Indeed Senior Vice President, ESG
LaFawn Davis, Indeed Senior Vice President, ESG

How Unconscious Bias Is Baked into AI

Davis began by sharing the shocking results of a recent AI experiment: 

  • When a generative AI tool was asked to generate images based on the prompt “working professionals,” the tool returned images of all white workers. 
  • When asked to generate images of “fast-food workers,” the tool returned images of people with darker skin tones 70% of the time, even though 70% of fast-food workers in the U.S. are white. 
  • When asked to generate images of “doctors,” women made up only 7% of the results, even though, according to the U.S. Bureau of Labor Statistics, 39% of doctors are women.

Clearly, AI systems reflect real-world biases. And as Davis argued, these biases have consequences. “The algorithm can determine whether you qualify for a loan, how you’re treated by the criminal justice system, or whether or not you score a job interview,” she said.

She cited the example of a multinational e-commerce company, whose hiring algorithm penalized job applicants that had earned degrees from historically women’s colleges. That’s because their machine learning models, she explained, were trained on historical data that skewed heavily male.

A large, seated audience at the FutureWorks conference.
At Indeed FutureWorks 2023, SVP of Environmental, Social and Governance LaFawn Davis talked bias in AI and how Indeed’s Responsible AI Team is making work better and more equitable for all.

Unlocking the Potential of AI to Break Down Bias and Barriers

Indeed has been exploring the positive potential applications of AI, along with its ethical implications, since 2020. In that time, the company developed an in-house Responsible AI Team.

“When you use Indeed’s matching and hiring platform, you’re using tools that were built with equity in mind,” Davis said. She added, “We’re on our own journey with equitable hiring, and are continuously evaluating and redesigning our hiring processes with a lens of inclusion. We’re adopting best practices as we learn — and we’re committed to driving change across fair chance hiring, accessibility, skills-based hiring and economic security.”

One area in which AI can level the field for job seekers and help employers find quality talent: higher-education requirements.

As Davis explained, 62% percent of Americans aged 25 and over don’t have college degrees. This statistic includes 75% of rural Americans, 72% of Black Americans and 79% of Hispanic Americans. Millions of job seekers are using search terms like “no college degree required” when looking for jobs on Indeed, according to Indeed data.

Hiring for skills has been proven to be five times more predictive of job performance than hiring for education, and more than two times more predictive than hiring for work experience alone. And yet, millions of Americans are automatically screened out of the recruiting process when an employer requires a college degree. 

Davis argued that millions of job seekers who are “skilled through alternative routes” are missing out — as are employers when looking for exceptional talent. Indeed’s Skill Connect makes it easier for job seekers to represent the skills and training they have received through hands-on technical and professional job training programs, and then matches them with the employers who need them.

Additionally, Indeed is developing AI tools to help recruiters “screen in” candidates, Davis said, detecting “skills in their resumes that may not have been initially considered.”

Indeed is exploring the potential of AI to assist job seekers in other ways, such as AI-driven chat-based search tools and AI-driven career coaches or career pilots. As mentioned in the morning keynote by Indeed CEO Chris Hyams, employers and recruiters who use Indeed’s AI Job Description Generator can save time for the more human aspects of hiring.

AI can also encourage pay transparency by reviewing wages and salaries at your company — and ensure, for example, that employers offer fair base compensation for new roles.

Keeping the ‘Human’ in ‘Human Resources’

Davis outlined a healthy list of all the things AI can do, before reminding the audience that it can’t do it all — nor should it. She emphasized the “human” in “human resources,” especially when it comes to determining hiring outcomes. Some 71% of Americans oppose the use of AI in making final hiring decisions. “Most U.S. workers don’t want robots having the final say,” she said.

Indeed has made a commitment to help 30 million job seekers who are facing barriers find jobs by 2030.

The key to shaping a better world of work alongside AI is to take action. “None of us knows with any certainty what the future holds, but if we all start incrementally building today toward a better tomorrow, we’re moving in the right direction,” Davis said. “When we work together, we have a greater chance of ensuring that job seekers facing bias and barriers aren’t left behind.”

Check out these other stories from Indeed FutureWorks 2023: