METR’s current frontier 50% task-completion horizon on a narrow suite of software tasks.
[10] Measurement caveat applies00 — Governing Question
The question underneath the technology.
This briefing does not prescribe a program. It puts a disciplined question before Black institutions: what authority will we retain over the terms of participation?
Will Black Americans enter the emerging technological order principally as users and workers — or also as builders, owners, operators, and governors?
Orientation, not advocacy · The case for informed agency
01 — The transition
Real capability. Incomplete reliability.
Artificial intelligence has moved from novelty to an institutional layer: connected to tools, files, code, databases, and business processes. Its value — and risk — comes from the systems into which it is embedded.
Organizations reporting AI use in at least one business function in Stanford’s 2026 AI Index.
[11] 2025 survey yearShare of notable frontier models produced by industry in 2025, according to Stanford.
[11] Industry concentration02 — Position
More than a jobs question.
Use and employment matter. They are not the full measure of agency. Technology distributes convenience, wages, contracts, intellectual property, equity, data, and governing authority to different actors.
STEAM, not merely STEM. E5 uses STEAM to keep arts, humanities, judgment, and civic purpose inside the technology conversation. The workforce figures below retain the National Science Board’s technical-workforce measurement and its broad classification: all education levels, including skilled technical occupations.
U.S. workers in the National Science Board’s broad technical workforce measurement in 2023 — one quarter of the workforce.
[15] 2023Black workers employed in those measured technical occupations, according to the 2026 National Science Board indicators.
[15] 2023Black-owned employer firms with $249 billion in receipts in 2023; 3.4% of U.S. employer businesses.
[16] 202303 — Lineage
Mastery, translation, institution-building.
Black participation in leading-edge technology is not a new accommodation to modernity. It is a recurring tradition of joining technical mastery to practical service, institutional capacity, and economic self-determination.
Institutional foundation
Booker T. Washington & Tuskegee
Build durable institutions, an economic base, and technical capacity rather than wait for outside systems to confer power.
Applied science & extension
Carver & the Jesup Wagon
A Tuskegee team carried practical demonstrations, equipment, samples, and instruction beyond campus — technology transfer before the phrase existed.
Capacity that survives an individual
Robert Russa Moton
Academic programs, endowment, physical plant, networks, Black professional governance — institution-building at scale.
Technical foundations
Langley, Dahlgren & Gladys West
Black mathematicians, engineers, programmers, and researchers helped make aerospace and GPS possible. Technical work can be invisible and still be foundational.
04 — Virginia
A living map of the question.
Virginia puts Black scientific history, leadership institutions, and globally significant digital infrastructure unusually close together. Select a point to reveal its strategic relevance.
The infrastructure beneath software is physical: data centers, chips, fiber, water, cooling, land, and electricity. U.S. data centers used about 4.4% of national electricity in 2023 and are projected to use roughly 6.7%–12% by 2028. The question is not whether buildout is simply good or bad. It is who gains knowledge, contracts, ownership, tax value, and governing influence — and who carries the costs.
05 — Agency
The ability to choose, govern, audit, and exit.
AI literacy is not merely knowing how to prompt a chatbot. It is the institutional ability to select a system, protect assets, evaluate performance, challenge outcomes, and leave without surrendering memory or control.
Purpose before tool
Begin with the institutional problem, the accountable owner, and the decision to improve — never vendor capability alone.
Data control
Know what enters the system, where it resides, whether it trains another model, and how it can be deleted or exported.
Human authority
Name the person responsible. High-consequence decisions cannot become unowned because automation contributed.
Evidence & traceability
Require sources, confidence labels, assumptions, and material-edit records. Persuasive language is not evidence.
Security & resilience
Treat models, connected tools, and uploaded documents as attack surfaces. Plan for prompt injection, leakage, and disruption.
Portability & exit
Preserve institutional memory outside any single vendor; retain data export, model substitution, and operational continuity.
Ownership & procurement
Track who captures revenue, IP, contract, equity, land value, and recurring margin — not only immediate convenience.
Talent & succession
Build enough internal competence to question, supervise, and retain judgment without dependence on one individual.
06 — Tensions
Responsible leaders refuse false choices.
Technology debates weaken when they collapse competing truths into slogans. The task is to hold the tensions together and design institutions capable of governing them.
Leadership question: Where can the institution move quickly, and where must it slow down because the cost of error is high?
Leadership question: Does access build internal capability, or make the institution dependent on a vendor it cannot audit or replace?
Leadership question: Which work should be automated, and which must remain deliberative because values, context, or accountability are central?
Leadership question: Are people prepared only to fill jobs — or also to form firms, own assets, win contracts, and set standards?
Leadership question: Who receives tax and economic benefits, and who bears land, water, grid, reliability, and ratepayer consequences?
07 — Questions
A conversation agenda, not a program agenda.
These questions belong to no single organization, ideology, or profession. They are for Black leadership across education, business, government, philanthropy, faith, law, and civil society.
What technical literacy will responsible institutional leadership require within the next three years?
Which decisions should never be delegated to an automated system, even when AI can assist analysis?
What data, IP, relationships, and institutional memory must remain under Black institutional control?
How can participation move beyond use and employment into contracting, company formation, equity, infrastructure ownership, and rule-making?
What procurement practices allow Black-owned firms to capture recurring value from the buildout rather than only temporary labor?
Which young leaders are being prepared to understand both the technology and the consequences of deploying it?
What does the Moton–Carver–West tradition suggest about the practical responsibilities of Black institutions in a new technical era?
How should institutions balance rapid experimentation with privacy, security, evidence, and moral accountability?
Method & sources
Seriousness begins with traceability.
Prepared with source verification current through July 26, 2026. Primary and official sources are preferred. Corporate, industry, and benchmark sources are treated as bounded evidence, not universal guarantees. Reference years are named where they differ from publication dates.
Document status: independent informational briefing; not commissioned by The Gloucester Institute; not a statement of Mrs. James’s or the Institute’s views; not a proposal, program, funding request, or solicitation.
[1–9]Historical record. The Gloucester Institute; National Park Service; Library of Congress; Tuskegee University; Smithsonian/USDA; Virginia State University; NASA.
[10–12]Capability & governance. METR Task-Completion Time Horizons; Stanford Institute for Human-Centered AI 2026 AI Index; NIST Generative AI Profile for the AI Risk Management Framework.
[13–14]Energy. U.S. Department of Energy Report on U.S. Data Center Energy Use; International Energy Agency, Energy and AI.
[15–16]Workforce & business. National Science Board / NCSES 2026 technical-workforce indicators; U.S. Census Bureau 2024 Annual Business Survey results for reference year 2023.
[17–20]Infrastructure & adjacent technology. Virginia JLARC Data Centers in Virginia; Virginia State Corporation Commission large-load rate class; NIST post-quantum standards; International Federation of Robotics Top 5 Global Robotics Trends 2026.