Exclusive: Researchers warn AI could add billions to U.S. racial wealth gap
Researchers warn generative AI could add $43 billion annually to America's already stark racial wealth gap over the next two decades, with Black workers also facing a higher risk of job loss thanks to automation.
Why it matters: The wealth gap between Black households and white households has widened since 1980 in the U.S., where the median white household now has $285,000 in wealth, compared to $45,000 for the median Black household.
Driving the news: Researchers from McKinsey conclude in a new report shared with Axios that AI threatens to automate half of all jobs that don't require a 4-year degree and that pay over $42,000 a year, which they define as the threshold for building wealth.
- The researchers concluded that the workers most at risk are in office support, production (e.g. machinists) and food services — all categories where Black workers are overrepresented and which many workers often use as stepping stones to higher-skill, higher-wage roles.
The big picture: Millions of Americans feel uneasy about their economic future.
- Rapid AI development risks feeding those fears unless it's accompanied by strong efforts to ensure the technology helps broaden prosperity.
Be smart: Remedying multi-generational wealth inequality is complex, and AI could compound the racial gap in several ways.
- The playing field is not level today: On average, Black Americans are paid less and they have less access to computers, broadband and digital skills than white Americans.
- Black workers are underrepresented in the tech sector developing AI.
- Black Americans face AI impacts ranging from a faster rate of job loss due to automation to unfair systems for accessing housing, credit and jobs due to racially-biased historical data and flawed AI risk management systems.
- One example of discrimination with multigenerational racial wealth impacts is the historical housing policy of "redlining."
Zoom in: The inequality trap described in the McKinsey report puts pressure on a wide range of parties to take action.
- The researchers recommend that Black workers protect themselves from AI disruption by learning skills that pay well and pose a lower risk of being automated.
- Occupations meeting these criteria that don't require 4-year degrees include dental hygienists, massage therapists and respiratory therapists.
- Employers in fields likely to automate first should contribute to reskilling efforts.
- AI developers should focus on insuring their models and algorithms aren't perpetuating or deepening biases.
- DEI and HR teams will need to quickly understand how to mitigate AI's negative impacts or risk becoming part of the problem.
Yes, but: AI can also assist economic mobility — for example, organizations could develop new generative-AI-based products aimed at rebalancing biases, avoiding discrimination, or targeting new products at underserved communities.
- Many AI proponents believe that the technology can level the educational playing field and bring better health care to underserved communities.
What they're saying: The report authors write that AI can yield equitable benefits, but only if developers "design generative AI with equity as a goal."
- "Gen AI will alter professional pathways that Black workers rely on to move from low-wage to higher-paying roles," they write.
- Report co-author Jan Shelly Brown, a McKinsey partner specializing in financial services, told Axios that she thinks "innovation in the AI ecosystem" offers the best hope for closing the wealth gap, but getting there means "more governance" to ensure AI trains on unbiased data.