America Is Building an AI Economy and National-Security System. Where Is Black Westchester?
34.7 percent. That was the error rate for darker-skinned women in one landmark study of commercial gender-classification technology. For lighter-skinned men, the maximum error rate was 0.8 percent. Another study examining pedestrian-detection technology found systems were about five percentage points less accurate at detecting darker-skinned pedestrians than lighter-skinned pedestrians. Then the National Institute of Standards and Technology evaluated facial-recognition algorithms using more than 18 million images of more than 8 million people and found demographic differences across the majority of algorithms studied. In some one-to-one systems, false-positive rates for African American and Asian faces were 10 to 100 times higher than for Caucasian faces.
Before somebody turns that into “Tesla couldn’t see Black people,” let me clean that up. The pedestrian research wasn’t specifically a Tesla study. The bigger issue is more important anyway: What happens when technology being taught to recognize the world doesn’t recognize everybody in that world equally? A computer doesn’t wake up racist, but an algorithm learns from data selected, labeled, tested and interpreted by human beings. When certain people are missing or underrepresented in that process, those gaps can show up in the technology.
That is why Black people need to pay attention to artificial intelligence. Not because AI is coming. It is already here. It is in our phones, schools, workplaces and businesses, and it is increasingly becoming part of the systems that make consequential decisions. Now the federal government is moving advanced AI deeper into America’s national-security apparatus, which means this conversation has officially moved way past ChatGPT.
On June 5, 2026, President Donald Trump signed National Security Presidential Memorandum 11, directing the federal government to accelerate AI adoption throughout the national-security enterprise. The directive addresses advanced commercial and open-source AI models, classified environments, high-security computing infrastructure, cybersecurity, military applications and the recruitment and training of an AI workforce. It also directs the Secretary of War to update federal policy governing autonomy in weapon systems as AI capabilities evolve, while maintaining human accountability through the constitutional chain of command. White House AI National Security Fact Sheet
Let that sink in for a minute. While some of us are still debating whether our children should use AI for homework, the United States government is discussing advanced AI on classified networks, high-security computing facilities, intelligence operations, cybersecurity, and policies governing autonomous weapons. That doesn’t mean AI has suddenly been permitted to independently decide whom to kill; the directive specifically emphasizes accountability and controllability. But it does tell us how far this technology has moved and how seriously the federal government is taking it.
This didn’t begin in June. The White House’s July 2025 America’s AI Action Plan established three major priorities: accelerating AI innovation, building American AI infrastructure, and leading internationally in AI diplomacy and security. The document talks about economic competitiveness, infrastructure, energy, manufacturing, workforce development, and America’s position in the global AI race. Americas-AI-Action-Plan.pdf. America isn’t sitting around trying to decide whether AI matters. America has a plan. My question is: What is ours?
For Black Westchester, this conversation needs to be bigger than whether you personally use ChatGPT. I want to know who is being trained, who is being hired, who is building companies, who is receiving contracts, and who owns pieces of the infrastructure supporting this new economy. I want to know what this means for Mount Vernon, Yonkers, New Rochelle, Peekskill, and White Plains while the opportunities are still developing, not after everybody else has positioned themselves and somebody suddenly discovers we weren’t in the room.
And please stop saying, “AI isn’t for me.” You don’t have to trust everything it tells you, and you absolutely should not. There are legitimate questions about privacy, misinformation, intellectual property, employment, and bias. But refusing to learn AI doesn’t prevent AI from affecting you. Your employer can adopt it. Your bank can use algorithms. Your child’s school will have to deal with it. Businesses can automate parts of their operations. Government can purchase AI systems. You don’t actually escape a technological transformation by refusing to open the app.
The federal government’s own Action Plan acknowledges that AI will transform work. It calls for expanding AI literacy and workforce skills, studying job creation, displacement, and wages, and developing rapid retraining and proactive upskilling for workers whose occupations may be affected. Americas-AI-Action-Plan.pdf. If Washington is preparing for jobs to change, our communities should not wait until that change reaches somebody’s paycheck before we start preparing.
Black people should be using AI, but use it with your brain turned on. If you own a beauty salon in Mount Vernon, explore how AI could help with marketing, customer retention, or understanding your business. If you’re a contractor in Yonkers, learn how it can help research opportunities and organize administrative work. If you run a nonprofit in New Rochelle, see how it can help organize program information and research funding. If somebody hands you a 90-page government document written like regular human beings were never supposed to understand it, let AI help break it down, then go back to the original document and verify what matters.
We should also be testing these systems on us. Ask about Black history, Puerto Rican history, Africa, the Caribbean, Mount Vernon and Yonkers. Ask about redlining, Black businesses, and the history of our communities. Ask questions about subjects you know well enough to recognize when something is missing or wrong. Our representation has to extend far beyond using a chatbot. We need Black people among the engineers, researchers, entrepreneurs, educators, evaluators, policymakers, investors, and owners determining what these technologies become. We don’t just need AI to recognize us. We need to be among the people deciding what AI recognizes.
Then follow the money, because artificial intelligence may feel like something floating around in “the cloud,” but that cloud needs a whole lot of stuff on the ground. AI requires data centers, electricity, cooling systems, semiconductors, telecommunications, cybersecurity, construction, and maintenance. The federal Action Plan specifically identifies electricians and advanced HVAC technicians among the occupations necessary to build and maintain AI infrastructure and calls for partnerships among government, employers, and workforce organizations to build training pipelines connected to employment.
Americas-AI-Action-Plan.pdf
Now we are having a different conversation. AI isn’t only an opportunity for the child who loves coding. It’s an opportunity connected to electrical work, cybersecurity, engineering, advanced manufacturing, construction, and infrastructure. Somebody is going to build these facilities. Somebody is going to cool them, wire them, secure them, and maintain them. Somebody’s company is going to receive those contracts. The question for Black Westchester is whether any of those somebodies are going to look like us.
The federal plan even calls for exposing middle- and high-school students to careers connected to AI infrastructure, expanding pre-apprenticeships, updating career and technical education, strengthening dual enrollment and expanding Registered Apprenticeships. Americas-AI-Action-Plan.pdf. So don’t just show our children how to prompt AI. Show them the server room. Show them cybersecurity, robotics, electrical infrastructure, intellectual property, and government procurement. And please show them what the invoice looks like. I don’t just want Black children impressed by technology. I want them paid by it, building it and owning pieces of it.
Westchester isn’t starting from zero, which makes the next question even more important. Businesses, colleges, schools, and community organizations are already beginning to engage with artificial intelligence. What I want to know now is who is getting through the door and what happens afterward. How many Black and Latino-owned businesses are accessing AI training? How many turn that knowledge into additional revenue? How many young people move from introductory programs into internships, apprenticeships, college programs, and employment? How many minority-owned companies eventually become vendors and contractors? Access isn’t a flyer, a panel, or putting “AI” in an event title. Access is what happens after somebody walks through the door.
We also need to talk about accountability. The June national-security directive says constitutional protections remain applicable and says AI must not be used for unauthorized or unlawful surveillance of Americans. Read the full National Security Presidential Memorandum. Those protections matter, and Black Americans don’t need conspiracy theories to understand why surveillance and accountability deserve scrutiny. Our documented history gives us enough reason to ask who authorizes these systems, what information they can analyze, how mistakes are corrected and who audits their outcomes.
Remember where this article started. Technology has already demonstrated demographic performance disparities. That doesn’t mean every AI system is racist or every algorithm is biased. It means testing matters, representation matters, and measurement matters. When technology begins influencing consequential decisions about human beings, you cannot correct what nobody is willing to measure.
This is the moment for Black Westchester to recognize what is happening while the architecture is still being built. America’s AI infrastructure is expanding. Workers are being trained. Businesses are developing products and services. Colleges are creating programs. Government policies are evolving, and advanced AI is now being integrated into the national-security enterprise. That means we still have time to position ourselves before this becomes another mature industry where we arrive asking why ownership doesn’t look like us.
I want the entrepreneur in Mount Vernon thinking about AI and revenue. I want the teenager in Yonkers understanding that an electrical apprenticeship could put them inside the infrastructure powering this technological transformation. I want the student in New Rochelle considering cybersecurity, engineering, or data science. I want the Black-owned company in White Plains watching procurement opportunities, and I want our schools, chambers, and workforce organizations making sure Peekskill, Mount Vernon, Yonkers, and the rest of our communities aren’t discovering these pathways after the best opportunities have already been claimed.
America has an AI Action Plan. The federal government is accelerating AI throughout national security. The infrastructure is being built, the workforce is being developed, and the money is beginning to move. Black Westchester cannot afford to arrive after the contracts have been awarded and ask why nobody called us.
AI is learning the world, and this time we need to recognize the moment. We need the technology to recognize us, the workforce to include us, the contracts to reach us, and when somebody pulls up the ownership records, some of those names better look like ours.
DON’T TAKE MY WORD FOR IT. RESEARCH IT. I want readers to read the documents, check the statistics, and come to their own conclusions. AI literacy also means source literacy. Open the reports. Read the federal policy. Question what I wrote. Question what the government wrote. Question what AI tells you. Then follow the information for yourself.
America’s AI Action Plan — The White House
White House Fact Sheet — AI in the National Security Enterprise
National Security Presidential Memorandum 11 — Full Directive
NIST — Study on Race, Age and Sex in Face Recognition Software
MIT Media Lab — Gender Shades: AI Accuracy and Skin-Tone Disparities
The MIT research is the source for the striking 34.7% error rate for darker-skinned women versus a maximum 0.8% for lighter-skinned men in the commercial gender-classification systems tested. NIST is the source for the much larger facial-recognition evaluation involving 18.27 million images of 8.49 million people and the finding that, in one-to-one matching, some algorithms had false-positive differentials of 10 to 100 times for Asian and African American faces compared with Caucasian faces.
