On Sept. 2, Mayor Zohran Mamdani announced a one-year moratorium on student-facing generative AI for nearly 600,000 New York City public school students, from 2K through eighth grade. Supporters call it protection. I understand the impulse, but I question the logic.
What, exactly, are we protecting Black children from — and what are we protecting them for? Beneath the calls for caution lies a familiar proposition: Black families should defend an educational order that has consistently failed to protect our children.
Consider New York’s 2024 Nation’s Report Card, which shows that 55% of Black fourth-graders scored below basic in reading and 56% of Black eighth-graders did so in mathematics. Federal civil rights data also document disproportionate discipline and unequal access to advanced coursework. Seven decades after Brown vs. Board of Education, our schools still sort Black children toward the bottom of outcomes we celebrate and the top of those we condemn.
Now we are told to exercise extraordinary caution because a new technology might compromise learning. Our children are already hurting. Why are we constantly asked to guard the gates of a house never built for us?
The stakes reach beyond classrooms. In September, Black unemployment stood at 7%, roughly twice the white rate. The median Black household has about $45,000 in wealth, compared with $285,000 for the median white household. McKinsey estimates that 24% of Black workers occupy jobs with high automation potential, compared with 20% of white workers. Automation does not guarantee displacement, but it will reshape opportunities for the next generation.
Read those numbers together, and the moratorium looks different. A community disproportionately exposed to technological disruption is being asked to delay its children’s preparation for that disruption. Without a serious alternative plan for equitable AI literacy, caution risks becoming a strategy for being left behind, drafted by people who will not be.
We have been here before. In 1999, the Commerce Department reported that Black households were only about two-fifths as likely as white households to have internet access at home. More than two decades later, Pew found that 80% of white adults had home broadband access, compared with only 71% of Black adults. That divide reflected unequal infrastructure, investment, and affordability — not a lack of Black curiosity.
Every new technology arrives with a chorus telling Black people to wait, and another chorus, years later, lamenting that we are behind. The cruelty is that these songs are often sung by the same singers.
Our children, meanwhile, are not waiting. Pew found that 31% of Black teens had used ChatGPT for schoolwork, compared with 22% of white teens. A school moratorium does not remove AI from children’s lives. It risks removing educators from the work of helping them understand it. Affluent families can purchase private tutoring, enrichment, and technological access. Children dependent on public institutions may be left with fewer opportunities.
The critics are right to worry about privacy, independent thinking, bias, and corporate influence. No one should hand unrestricted chatbots to preschoolers or confuse screen time with learning — but these concerns demand thoughtful design, not an absence of preparation.
Research offers reasons for both caution and possibility. One study found that unrestricted AI assistance could undermine students’ independent performance, while instructional safeguards reduced that harm. Other trials of carefully designed AI tutoring have reported substantial learning gains in specific settings. These findings are not a blank check for AI. They show why design, educator guidance, and independent evaluation matter.
To be fair, New York City’s policy is not a total ban on AI education. It includes high school pilots, literacy instruction for older students, and certain exceptions. Those provisions matter, but they do not explain why younger students cannot receive developmentally appropriate instruction about AI — or how the city will prevent a new divide between children whose families can afford technological opportunities and those who cannot.
We should be preparing educators, teaching students to question AI’s answers and biases; protecting their data; and inviting Black families and communities into decisions about which tools are built, purchased, and governed. We should ask whose knowledge is represented, who profits, who bears the risks, and who gets to define educational progress.
If Black people do not help define what AI means for our classrooms, clinics, and storefronts, others will define it for us. We must refuse the fantasy that technological progress automatically produces human progress. We also must refuse the notion that our safest place is on the sidelines of history.
Let us be cautious without being absent. Let us protect our children without surrendering their possibilities. We have spent too many generations living in futures authored by others. This time, let’s write our own.
David E. Kirkland, Ph.D., is an award-winning scholar; author of Pedagogy of the Black Child; former NYU professor; and founder and CEO of forwardED, a national education research and consulting firm advancing equity, educational transformation, and human agency.
