The impact of generative AI on Black communities

Generative AI (gen AI) has already initiated a seismic shift in work and value creation. A recent McKinsey report, The economic potential of generative AI: The next productivity frontier, identified up to $4.4 trillion in potential global economic impact from gen AI across functions and industries. With gen AI in its infancy, organizations are just beginning to understand the potential of applying it to their own goals.

As often happens, the advent of a new technology can create or exacerbate divides, including the racial wealth gap. This article explores how gen AI may affect Black communities and Black workers. While the technology has the potential to widen the racial gap, it’s also possible to shepherd this new technology in an inclusive way that not only yields equitable benefits for Black and other marginalized communities, but also accelerates the closure of the racial gap.

Gen AI’s impact on Black economic mobility

New wealth created by digital and AI capabilities flows through an economy where the median Black household has only about 15 percent of the wealth held by the median White household. The data are striking: the median Black family has amassed about $44,900 in wealth; the median White household holds $285,000 in total assets.

Without correcting long-standing patterns, gen AI has the potential to increase this racial wealth gap. Annual global wealth creation from gen AI is projected to be about $7 trillion, with almost $2 trillion of it expected to go to the United States, given its share of global GDP. US household wealth captures about 30 percent of US GDP, suggesting the United States could gain nearly $500 billion in household wealth from gen AI value creation. This increase would translate to an average of $3,400 in new wealth for each of the projected 143.4 million US households in 2045.

Black Americans capture only about 38 cents of every dollar of new household wealth despite representing 13 percent of the US population. If this trend continues and projections of the growth of Black households are accurate, by 2045, racially disparate distribution of new wealth created by gen AI could increase the wealth gap between Black and White households by $43 billion annually (Exhibit 1).

By 2045, skewed distribution of annual gen AI wealth creation could cost Black households in the United States $43 billion each year.

Gen AI’s impact on Black workers

One of the most pressing questions about gen AI is how it will affect workers and, specifically, workers of color. Black Americans, with an unemployment rate twice that of White workers, are overrepresented in roles most likely to be taken over by automation. Indeed, Black workers are increasingly overrepresented in four of the top five occupations at risk of automation: office support, production work, food services, and mechanical installation and repair. If reskilling efforts are not undertaken, this trend only stands to worsen. According to our analysis of 2022 data, some 24 percent of all Black workers are in occupations with greater than 75 percent automation potential, compared with just 20 percent of White workers. It’s no wonder, then, that in a recent survey from workplace research firm Charter Works, 53 percent of Black respondents said they worry that AI might replace their jobs in the next five years, compared with 39 percent of White respondents.

But those figures alone don’t tell the full story of how gen AI may impact Black workers.

Gen AI’s impact on high-mobility jobs

Gen AI does not put low-wage jobs at particular risk in the same way previous advances in automation did—it actually has greater automation potential for higher-wage knowledge work. However, gen AI will likely alter professional pathways that Black workers rely on to move from low-wage to higher-paying roles.

Understanding high-mobility jobs

Inclusive, high-mobility jobs are those that provide livable wages and potential for an upward career trajectory and development over time but do not require a four-year college degree. High-mobility jobs have two tiers:

  • “Gateway” jobs, based on experience rather than degrees, offer a salary of more than $42,000 per year and can unlock a trajectory for career upward mobility.
  • “Target” jobs, the next level up for people without degrees, are attractive occupations in risk and income, offering generally higher annual salaries and stable positions.

Seventy-four percent of Black workers do not have college degrees, but in the past five years, one in every eight has moved to a gateway or target job. Gen AI may significantly affect those occupations, as many of the tasks associated with them—including customer support, production supervision, and office support—are precisely what gen AI can do well. Coding bootcamps and trainings have risen in popularity and have unlocked access to high-paying jobs for many workers without college degrees. But such pathways are also at risk of disruption, as gen AI–enabled programming has the potential to automate many entry-level coding positions.

Ultimately, between 2030 and 2060, gen AI may be able to perform about half of the gateway or target jobs that many workers without degrees have pursued—closing a pathway to upward mobility that many Black workers have relied on (Exhibit 2).

Half of all ‘gateway’ or ‘target’ jobs could eventually become fully automated.

‘Future proof’ has a new paradigm

Previous assumptions about channeling workers without degrees toward paths with a growth trajectory held that “future-proof jobs,” such as coding, would become increasingly important. Gen AI, with its ability to create outputs that were previously believed to require human work (including coding and the development of some creative content), will change this paradigm. The rapidly expanding scope of what gen AI can do—both now and in the near future—means that workers must pivot from future-proof jobs to “future-proof skills” (Exhibit 3).

‘Future-proof skills’ can help workers avoid being displaced by future waves of automation.

These skills include socioemotional abilities that require a high degree of emotional intelligence (for example, talk therapist, communications manager), a physical presence with hand-eye coordination that’s difficult for machines to replicate (for instance, physical therapist), or a comfort with ambiguity and the ability to engage in nuanced problem solving in an array of specific contexts (for example, senior executives). In the tech industry, the distinction between future-proof jobs and future-proof skills may be in, on the one hand, knowing merely how to use a specific coding language to code formulaically versus, on the other hand, understanding computational and statistical principles in order to troubleshoot and solve problems with machine-generated code. Although gen AI can create first drafts of code, only software engineering managers have the nuanced skills to test and edit all kinds of code and to coordinate across multiple teams.

Focusing efforts on developing nonautomatable skills such as these will better position Black workers to develop the increased resilience needed to weather the rapid changes that gen AI will bring.

Gen AI’s potential to accelerate the eight pillars of Black economic mobility

A recent report from the McKinsey Institute for Black Economic Mobility identified eight pillars that represent areas with the highest potential to move the needle on Black economic mobility. They are financial inclusion, credit and ecosystem development for small businesses, health, workforce and jobs, pre-K–12 education, the digital divide, affordable housing, and public infrastructure.

When gen AI meets these eight pillars, there’s potential for profoundly different levels of impact—both positive and negative—depending on how gen AI tools are trained, designed, adopted, and used (Exhibit 4).

Generative AI poses both opportunities and risks for Black economic mobility.

Gen AI has the potential to influence the eight pillars positively or negatively. The next section delves into the opportunities for supporting more equitable outcomes for Black communities in two of the pillars: healthcare and financial inclusion.

Healthcare

In healthcare, gen AI has the potential to address the inequitable access to care that has historically impeded Black Americans from leading healthy, long lives. Twenty-six percent of Black Americans do not have a stable healthcare provider. According to a 2021 study from the University of Southern California, half of the 15 million Americans who live in pharmacy deserts are Black or Latino. And according to the Kaiser Family Foundation, Black patients fared worse than White patients across 75 percent of health measures (including life expectancy, chronic diseases, cancer incidence, and mortality). Gen AI may offer an intervention in some of these scenarios.

One of the most striking examples of racially disparate healthcare outcomes is the racialized inequality and inequity for Black mothers and children, both during and immediately after pregnancy. Across the US, Black pregnant people face significant disparities in outcomes. Nine percent of pregnant Black Americans receive delayed or no preterm care. More than 14 percent of Black births are preterm, which can have lasting health effects through childhood and adulthood.

However, incorporating AI into prenatal care has reduced preterm births. In one study described by Harvard Business Review, AI was used to support pregnant people throughout their care, including by aggregating and synthesizing patients’ medical histories (including any history of preterm deliveries or preeclampsia) and social determinants of health that affect pregnancies (such as food insecurity or lack of access to transportation). This AI-driven support enabled providers to more easily identify a range of risk factors earlier in patients’ pregnancies, and to connect those patients to the right care sooner than they might have done otherwise.

Gen AI has the potential to increase the power of such early interventions by, for example, drafting personalized summaries and after-visit action plans in plain language for providers and patients to reference. If included in the care of all Black pregnant people, this technology could help thousands of Black babies be born at term, instead of preterm, every year.

Using gen AI in supportive applications could help in achieving the following healthcare goals:

  • Using electronic medical records to address unmet medical needs. Expediting and synthesizing patient histories to improve clinical decision making helps providers connect patients with the care they need.
  • Increasing patient adherence to care protocols. By analyzing patient data and predicting when patients may stop taking their medications, healthcare providers can intervene by, for example, administering reminders and personalized dosing instructions.
  • Improving signal detection of adverse-event reports. Natural-language summarization can assist in vigilance about medication use generally, and for specific patient communities.
  • Providing personal assistance to patients. On-demand virtual personal physicians or pharmacists equipped with a patient’s data and medical history can serve as a first point of contact.
  • Predicting and creating next-best-channel content. Using AI to understand the channel and tech preferences of individual patients and patient communities can drive increasingly bespoke outreach and provide tailored support in how patients receive information and care.

Financial inclusion

Applications of gen AI also have a chance to increase equity by enhancing access to banking products for Black Americans, who have historically faced exclusion from financial products and services. Analysis from the McKinsey Institute for Black Economic Mobility shows, for example, that Black consumers have had less access to physical banks and ATMs. Counties with populations that are majority people of color have roughly 30 percent fewer banks than majority White counties. Black consumers also have fewer cost-effective options for loans. Car shoppers of color are 63 percent more likely to be offered more expensive loan options than White car shoppers with similar or lower loan qualification scores. And in general, Black consumers are twice as likely as White consumers to be denied credit.

One aspect of historical racism’s legacy in consumer banking is that Black consumers must often rely on nonbank financial services that are costly and even predatory. Federal Reserve data show that 40 percent of Black American households are currently considered to be either unbanked (they do not use or have access to a checking or savings account) or underbanked (they have financial accounts but still rely primarily on alternative financial services, such as money orders and foreign remittances.) Only 12 percent of White households fall into those categories. Underbanked and unbanked households of all races pay $173 billion annually in fees and interest annually to nonbank financial services (such as payday loans, check cashing facilities, money orders, and pawning). Adequate levels of banking for Black consumers could avoid the high costs of those nonbank services. If 25 percent of underbanked Black adults became adequately banked by 2025, they would collectively save $23 billion in annual fees, enough to make down payments on two million homes. The financial value to Black communities of increasing access to traditional consumer banking services and products is considerable. In fact, more than 50 percent of Black adults say they want to learn more about financial products available to them.

Black consumers are also increasingly drawn to virtual banking options: 14 percent of Black adults consider a digital bank to be their “primary” banking provider, compared with 8 percent of White adults.

Gen AI has the potential to play a strong role in connecting more Black consumers with traditional banking services, by providing specialized, bespoke, and accessible services to those consumers.  For example, gen AI could be used to create targeted marketing that responds to Black consumers’ needs, interpret clients’ inquiries and create specific financial plans for saving, and, ultimately, help those consumers grow savings through investment and other wealth-development strategies.

Potential supportive applications of gen AI could help to achieve several financial inclusion goals:

  • Increasing accessibility to banking by generating targeted, responsive marketing content and specific, personalized product offerings.
  • Supporting the creation of long-term, bespoke financial plans via enhanced interpretation of client inquiries based on customer and account information.
  • Enhancing compliance monitoring. With AI policy bots and other monitoring systems, equitable and transparent access to financial products can be ensured.

Ultimately, gen AI alone will not reduce or increase inequality. The technology must be applied so that it increases access for Black patients, customers, and communities in ways that thoughtfully address racial gaps along the eight pillars, with interventions aimed at closing those gaps. Leaders with influence in each of the pillars—including healthcare providers, commercial banks, educators, and CEOs—must build an equity lens into their developing gen AI strategies to ensure that their organizations actively seek to increase equitable outcomes—with positive effects for Black and other marginalized communities as a priority, not an afterthought.

Deploying gen AI with an equity lens

Leaders are at a key juncture with gen AI: they can apply it in a manner that promotes inclusion, fairness, and opportunity, or they can let it evolve without particular attention to equity and societal implications. Leaders developing gen AI strategies for their organizations can consider how to use their influence to build awareness of the technology’s potential impact on equity in the following ways:

  • Vigilant implementation. Prepare workers to meet the needs of a “post–gen AI” landscape, and use gen AI with care.
    • Reskilling workers. Workers at risk of replacement by gen AI can be trained in skills that cannot be easily displaced by technology in the near future and that are transferable across roles. That means emphasizing foundational and nuanced skill-building instead of role-specific training.
    • Judicious use. All leaders and organizations implementing gen AI must ensure that it is used in contexts where it can dependably and fairly make decisions—not where understanding nuance or cultural, historical, or social factors is important, or where complex, high-risk questions must be framed and resolved.

  • Sustaining responsible gen AI. Establish guardrails to ensure that as this technology develops, people and communities have an active role in shaping how it is and isn’t used.
    • Regulation. Lawmakers and regulators should continue to understand new and newly common applications of gen AI across industries and functions, to ensure that the technology is not used in a way that impairs Black and other marginalized communities. They should also consider establishing “rules of the road,” so that companies using gen AI are protecting sensitive data and information from being compromised and misused.
    • Democratized access. Gen AI should be equally accessible to all. The technology will flourish if its user base is broad enough to enable new, creative, important implementations—and if it avoids creating another digital divide between Americans who have access to it and those who do not. Moreover, digital literacy must now include gen AI and be widespread across educational curricula for all students.

  • Building responsible gen AI. Cultivate conditions that allow key inputs to successful gen AI to be fair, unbiased, and representative of diverse communities of people and their data.
    • Nutritious data. Training foundation models on data sets that are unbiased, authentically representative, and free from racial, ethnic, gender, or other biases will help to ensure that gen AI products create fair, accurate, reliable outputs.
    • Diverse tech talent. Advancing and growing the bench of Black and other underrepresented tech talent will ensure that gen AI is developed and shepherded into the future by a wide talent base with varied experiences, perspectives, and bases for understanding the impact and potential of gen AI and other technologies.
    • Participatory design. All stakeholders of gen AI applications—including end users and those affected by new uses—should participate in the design of new gen AI products.

Gen AI could fundamentally change how most work is done and most services are provided. Early in gen AI’s trajectory, leaders have an opportunity to build equity and fairness into their developing gen AI strategies and associated applications, and to accelerate the closing of the racial gap. The aim: to have all people and communities—including Black communities, previously left behind by seismic shifts in tech—benefit from this amazing technology transformation. The time to act is now, using the foresight of past transformations as a guide.