AI bias is an emerging hurdle for DEI leaders, expert cautions

As AI is embedded into most HR functions, some inclusion experts have publicly voiced concern that it could bring with it new biases. But it’s not just the AI tools that leaders should concern themselves with. It’s every step of the development process.

Nicol Turner Lee, senior fellow and governance studies director at the Center for Technology Innovation at the Brookings Institution, recently addressed bias in AI in a room of lawyers and diversity and inclusion practitioners at the Meltzer Center for Diversity, Inclusion and Belonging at New York University School of Law.

“There are so many people, even to this day, who are actually not part of the digital ecosystem,” just one issue with the development process, she said. Therefore, “these technologies are not necessarily designed by us…These are actually technologies being created and controlled by a handful of companies that are dictating the direction and where AI shows up.”

Digital invisibility, she said, doesn’t just apply to those who lack broadband access, but those who are ignored in AI development and by AI systems. (For example, disability advocates have expressed concern that mainstream AI image generators do not accurately depict disabilities.)

And while much attention has been paid to AI tools, there’s been less focus on deep learning, a type of machine learning that underlies AI tools, like pre-hiring algorithms that rate pre-recorded job applicant video interviews, according to Turner Lee.

“That video shows whether or not that person smiles often, has great eye contact,” she said. “That same application really omits and excludes Black and Latina applicants because they’re perceived as not being friendly based on the emotional cognition of their face.”

AI has made work and life easier in many ways, but Turner Lee noted that “AI bias is real.” So it’s important that leaders consider who is at the table and who develops and chooses AI, because biases are more likely to show up when diversity isn’t incorporated from the beginning.

“This is an issue that you’ve got to care about,” she concluded. “I know you spend a lot of time making sure that the role that you were charged to do to advance inclusion in your organization still stays steadfast, and I also know that you all have had the courage to continue this battle. But I want to suggest to you: It gets more complicated with AI.”