AI is expected to generate roughly $500 billion in new household wealth in the U.S. by 2045, according to a December report from consulting firm McKinsey & Co. But considering the racial wealth disparities that already exist, in which Black households hold roughly 15% of the wealth of the median white household, the gap between Black and white households could grow to $43 billion annually, according to McKinsey.

Jan Shelly, a partner at McKinsey and lead author on the report, says this kind of racial wealth gap widening is not inevitable, but avoiding it will require companies and leaders to intervene as they plot AI’s path forward.

“AI is going to be prolific, everyone knows that. What I don’t think anyone anticipated was the hockey stick kind of impact it was going to have through the introduction of ChatGPT,” Shelly says. “What we don’t want now is the erosion of the gains we’ve had in Black communities.”

One of the biggest areas for concern among Black workers, Shelly says, is AI’s impact on high-mobility jobs, or those typically knowledge-worker roles that offer livable wages and opportunities for earning more money without requiring a four-year degree. 

Around 74% of Black workers do not have a college degree, but one-eighth of those workers have sought out high-mobility jobs over the past five years, according to McKinsey’s report. Those jobs can be categorized into two buckets: gateway jobs, which are solid, well-paying stepping stone roles, and target jobs—higher profile jobs in attractive professions. Both categories, Shelly says, will experience a high rate of disruption.

Between 2030 and 2060, gen AI could be able to perform roughly half the gateway or target jobs of workers without degrees. But Black workers have typically relied more on gateway jobs to gain a foothold in higher paying professions, according to McKinsey, making them particularly vulnerable to AI disruption. 

So what needs to be done to intervene? First and foremost, companies need to “bring their chief diversity officers to the table” for decisions regarding AI use cases and talent implementation, Shelly says. “That needs to be the base standard. That would make me hopeful.”

AI should also be seen as a job augmenter, not a job killer. The technology will be able to assist some of the more mundane tasks of lawyers and paralegals, for example, but that doesn’t mean those jobs are no longer necessary. 

Organizations need to begin preparing their workforce without leaving segments of it behind, according to Shelly. This includes not only upskilling for things like how to properly code or prompt language models, but going one step further in training workers to gain the necessary knowledge to troubleshoot and solve problems in machine-generated code, a more nuanced task AI isn’t good at.

Companies should also be thinking about how to hire and train for roles that require non-automatable socioemotional skills that call for a high level of EQ. This includes a comfort with ambiguity and being able to do nuanced thinking and problem solving. Shelly says she hopes to see organizations do more work to instill these kinds of skills in K-12 and college education to better prepare future Black workers.

It’s no secret, however, that corporate America has not been particularly good at marrying DEI and talent strategies. But there has been some progress, Shelly says, and the introduction of AI has the ability to level the playing field a bit more for Black workers. 

“It gives me hope that organizations are already thinking about what are the right communications and processes when it comes to training their workforce to use gen AI,” Shelly says. “It can be the great leveler, but it can exacerbate the gap as well. At the end of the day, the wealth gap isn’t going to close itself without intervention.”