Before Carter G. Woodson created Negro History Week, he worked in coal mines.
This is not the part of his biography that gets repeated in February. We remember the Harvard doctorate, the Association, the journal, the week that became a month. We remember the intellectual. We forget the laborer. We forget that the man who built the infrastructure of Black historical memory spent years underground, in the dark, doing the work that the economy assigned to people who looked like him.
It matters because Woodson understood something about economic life that his academic peers sometimes missed: literacy was not a luxury. It was an economic instrument. In the coal mines of West Virginia, Woodson read aloud to formerly enslaved men who could not read themselves — men who carried knowledge, experience, and intelligence that the system refused to credential. He saw, up close, the gap between capacity and access. And he spent the rest of his life building the bridge across it.
One hundred years later, that gap has a new name. It is not called illiteracy. It is called the digital divide, or the AI skills gap, or the automation displacement risk. The language is clinical. The stakes are not. What is being decided right now — in boardrooms, in legislatures, in venture capital meetings, in school district budgets — is whether artificial intelligence will become the next instrument of Black economic exclusion, or whether it will be designed, distributed, and governed in ways that break that pattern for the first time.
This is not a technology question. This is a history question. And the history is unambiguous: every major shift in Black economic life has been determined by who controls access to the dominant literacy of the era.
Literacy Has Always Been the Gate
The economics of Black America cannot be separated from the economics of knowledge access. This is not metaphor. It is mechanism.
In the antebellum South, anti-literacy laws were not cultural preferences. They were economic policy. An enslaved person who could read could forge a pass, understand a contract, interpret a deed, follow a newspaper, comprehend the law that governed her bondage — and, critically, communicate that understanding to others. Literacy was dangerous not because reading was inherently revolutionary, but because reading was economically revolutionary. It disrupted the information asymmetry on which the entire plantation economy depended.
When slavery ended, the formerly enslaved understood this immediately. The explosion of freedmen’s schools during Reconstruction was not sentimental. It was strategic. Within five years of Emancipation, thousands of schools had been established across the South — most of them created not by Northern missionaries or the Freedmen’s Bureau, but by Black communities themselves, pooling what little they had to hire teachers, build schoolhouses, and fill them with children and adults alike. By 1870, the Bureau alone was supporting over four thousand schools educating a quarter of a million students. Researchers have since demonstrated that Black children exposed to greater educational opportunity during Reconstruction saw significantly better economic outcomes as adults — and that those benefits transferred to their children. Literacy was not just personal advancement. It was intergenerational economic infrastructure.
Then came the backlash. Jim Crow did not merely segregate schools. It *defunded* them. It redirected tax dollars from Black education to white education. It ensured that the literacy gap Reconstruction had begun to close would be forcibly reopened and maintained for decades. The economics were never incidental to the racism. They were the point. A people denied access to the dominant literacy of their era — whether that literacy is reading, computing, or, now, AI fluency — are a people whose economic participation can be controlled by those who possess it.
The GI Bill repeated the pattern. On paper, it was race-neutral. In practice, its benefits flowed overwhelmingly to white veterans. Black veterans were steered toward vocational programs rather than four-year universities, denied mortgages in suburban developments, excluded from the wealth-building mechanisms that created the white middle class. The credential economy that emerged in the postwar period — in which a college degree became the primary gateway to middle-income employment — was built during a period when Black access to that credential was systematically constrained.
The digital economy followed the same script. When computing became the dominant economic literacy of the late 20th century, Black communities faced the “digital divide” — a phrase that sounds quaint now but described a real and consequential gap in access to hardware, broadband, training, and the institutional networks that translated digital skills into economic opportunity. The gap was not natural. It was produced — by decades of disinvestment in Black schools, by the geographic concentration of tech industries in overwhelmingly white enclaves, by venture capital structures that directed funding away from Black founders at rates that remain staggering.
Each of these moments follows the same structure: a new form of literacy emerges. That literacy becomes the gateway to economic participation. Access to that literacy is distributed unequally along racial lines. The gap compounds. The people locked out are told to catch up. And by the time they do, the gate has moved.
AI is the new gate. The question is whether the pattern holds.
The Numbers Are Not Neutral
Let’s be direct about what is at stake.
McKinsey’s research on generative AI and Black economic mobility presents a stark finding: if the wealth created by AI is distributed along current racial lines, the gap between Black and white household wealth could grow by $43 billion annually by 2045. That is not a projection of decline. It is a projection of *acceleration* — a widening that compounds on top of centuries of existing disparity, driven by a technology that is supposed to represent progress.
The mechanisms are precise. Black workers are overrepresented in four of the five occupational categories most vulnerable to AI-driven automation: office support, production work, food services, and mechanical installation and repair. These are not peripheral jobs. They are the entry points — the first rungs of the ladder that millions of Black workers have used to reach middle-income stability. When those rungs are automated away, the question is not just whether new jobs will be created. It is whether Black workers will have access to the training, credentials, networks, and capital required to reach them.
Meanwhile, the high-growth sectors of the AI economy — machine learning engineering, data science, AI product management, AI governance — remain overwhelmingly white and male. Black representation in the tech workforce has barely shifted in a decade. Funding to Black founders in AI and adjacent fields has declined for three consecutive years. The pipeline programs that are supposed to address these gaps remain chronically underfunded and structurally disconnected from the industries they are meant to feed.
And yet. The picture is not uniformly bleak. Recent research from Jobs for the Future found that Black workers and learners are not merely keeping pace with AI adoption — they are in many cases leading it. Over half of Black survey respondents reported using AI tools daily or weekly, significantly outpacing the national average. Black workers were more likely to report that AI was already changing their jobs and more likely to be experimenting with AI tools on their own initiative rather than under employer direction.
This is a critical finding, and it complicates the narrative of passive victimhood that too often accompanies discussions of Black communities and technology. Black people are not waiting to be told about AI. They are using it. The gap is not in interest or initiative. It is in the *infrastructure* — the training programs, the time to experiment, the institutional support, the professional networks, and the capital that turn individual fluency into collective economic power.
AI Literacy Is Not Computer Science
Here is where the conversation typically goes wrong.
When policymakers discuss AI and economic equity, they almost invariably reach for the same prescription: more STEM education, more coding bootcamps, more computer science degrees. These are not bad ideas. They are incomplete ideas. And their incompleteness is itself a form of gatekeeping.
AI literacy in 2026 does not mean everyone needs to become a machine learning engineer. It means something more fundamental and more democratic: the ability to understand what AI systems are, how they make decisions, where they are operating in your life, and how to negotiate with them — as a worker, a consumer, a citizen, a business owner, a patient, a parent.
This is closer to financial literacy than to computer science. And that analogy matters, because the history of financial literacy in Black America illustrates both the necessity and the insufficiency of individual education alone.
For decades, the financial literacy movement told Black communities that the path to wealth was knowledge: learn to budget, learn to invest, learn to manage credit. This was not wrong. But it was radically incomplete, because it placed the burden of navigating a structurally hostile financial system entirely on the individuals being harmed by it. It did not address the predatory lending, the discriminatory appraisal practices, the redlined mortgage markets, the underfunded community development institutions — the architecture that made financial knowledge necessary in the first place.
AI literacy must avoid this trap. Teaching people to use AI tools is necessary. But it is not sufficient if the tools themselves are designed without their input, if the data those tools are trained on carries historical bias, if the industries deploying those tools are not hiring from their communities, and if the governance structures overseeing those tools do not include their voices. Literacy without structural power is a treadmill. You run faster. The gap stays the same.
What Black communities need is not just AI education. They need AI *infrastructure* — the combination of skills, access, institutional support, and governance participation that transforms individual capability into collective economic agency.
Design Choices Are Moral Choices
The AI industry likes to present its products as inevitabilities. The technology is coming. Disruption is unavoidable. The future is automated. This framing serves a purpose: it positions the companies building these systems as forces of nature rather than actors making choices.
But every AI system is a series of design choices. Who is the training data sourced from? Whose labor annotates it? What objectives is the model optimized for? Who tests it before deployment? Who is displaced when it is deployed? Who profits? Who decides?
These are not technical questions with technical answers. They are moral questions with economic consequences. And for Black communities, the answers to these questions will determine whether AI becomes another extraction engine or something genuinely different.
Consider hiring algorithms. A company deploys an AI tool to screen resumes. The tool is trained on historical hiring data. That data reflects decades of discriminatory hiring practices. The tool learns the patterns. It reproduces them. It does so at scale, at speed, and with a veneer of objectivity that makes the discrimination harder to identify and harder to challenge. This is not a hypothetical. It has happened. It is happening. And unless the design choices change, it will continue to happen.
Now consider the alternative. An AI tool designed in partnership with workforce development organizations that serve Black communities. Trained on data that has been audited for bias. Deployed with transparency about how it makes decisions. Governed by a board that includes the people most affected by its outcomes. Measured not just by efficiency but by equity. This tool is also possible. It is not less technologically sophisticated. It is more morally sophisticated. And it requires that the people building these systems make choices that prioritize inclusion over speed and accountability over scale.
The difference between these two tools is not a matter of technology. It is a matter of will.
What Must Be Built
Resisting both the fantasy that AI will automatically liberate Black communities and the fatalism that it will inevitably harm them, SOLE insists on a third position: that the impact of AI on Black economic life is not predetermined. It is designed. And what is designed can be designed differently — but only if Black communities are present at the design table, not as consultants, but as architects.
This requires investment in four areas:
Community-based AI literacy at scale. Not coding bootcamps for the few. Accessible, practical education that helps working people understand how AI is already shaping their jobs, their credit, their healthcare, their children’s education — and how to advocate for themselves within those systems. This must be embedded in the institutions Black communities already trust: churches, barbershops, community colleges, unions, civic organizations. It must meet people where they are, not where the tech industry wishes they were.
Black-led AI research and development. The people building AI systems must include the people most affected by them. This means sustained investment in Black AI researchers, Black-led technology firms, and AI research centers housed within HBCUs and other institutions rooted in Black intellectual tradition. Not as diversity initiatives. As economic strategy.
Workforce transition infrastructure. For the millions of Black workers in roles at high risk of automation, the question is not whether the transition is coming but whether it will be managed with dignity. This means federally and state-funded reskilling programs designed with — not for — Black workers. It means income support during transition. It means portable benefits that do not evaporate when a job is automated. It means treating displaced workers as assets to be invested in, not problems to be managed.
Governance participation. Black communities must be present in the rooms where AI policy is made — in corporate AI ethics boards, in federal regulatory processes, in standards-setting bodies, in the venture capital firms that decide which AI products get funded. Economic power in the AI age will belong to those who govern the systems, not just those who use them.
The Bridge That Refuses to Break
Woodson read aloud to men in coal mines. He did not do this because he believed reading would save them from the mines. He did it because he understood that knowledge — shared, collective, accessible — was the precondition for any economic transformation that might follow. The literacy came first. The agency came second. The institutions came third. And the economic change, when it came, came because people had built the infrastructure to demand it and the capacity to sustain it.
One hundred years later, the literacy has changed. The infrastructure has changed. The technology has changed. But the pattern has not.
Black economic mobility in the AI age will not be determined by whether the technology is powerful. It will be determined by whether the people most affected by it have the literacy to understand it, the access to use it, the institutions to govern it, and the collective power to insist that it serve their interests — not just as workers, not just as consumers, but as architects of the systems that will define economic life for the next century.
Woodson built the bridge between knowledge and power in 1926. That bridge has held for a hundred years, despite every attempt to burn it down.
The question for 2026 is whether we will build the next one.
