Generative AI is being promoted as the next great equalizer in American education, with Microsoft CEO Satya Nadella, investor Marc Andreessen, and a growing number of policymakers arguing it will give every student a personalized tutor, writing coach, research assistant, and study partner available around the clock. A historian of education and former public school teacher who has spent 35 years watching education reforms says this vision repeats a familiar pattern: schools alone have never closed achievement gaps.
Each previous reform promised to transform opportunities for disadvantaged students. Each produced some genuine benefits. None fundamentally changed the inequalities that shape educational achievement in the U.S. That's because Americans have long asked schools to solve problems that originate far beyond the classroom.
The pattern of reform
The belief that schools can equalize opportunity is almost as old as public education itself. In 1848, Horace Mann, the first secretary of the Massachusetts Board of Education, called schools "the great equalizer of the conditions of men." Nearly two centuries later, that ideal still shapes how Americans think about solving inequality.
When achievement gaps persist, the response is rarely to ask whether schools alone can overcome them. Instead, the search begins for the next educational breakthrough.
In the 1990s and 2000s, that breakthrough was charter schools. Advocates argued competition and innovation would dramatically improve outcomes for students in poor and underserved communities. Some charter schools did. Many did not. A 2023 Stanford University study found modest academic gains overall, but nothing approaching the transformation supporters had promised.
Around the same time, Teach For America recruited graduates from elite universities to teach in under-resourced schools, hoping young, talented teachers could overcome educational disadvantage. Many became excellent educators. But high turnover limited the program's reach, and no teaching corps, however committed, could erase the effects of poverty, housing instability, or unequal access to opportunity.
Then came No Child Left Behind. Signed into law in 2002, it promised that standards, testing, and accountability would finally close gaps in student achievement across race and class. The law exposed disparities, but it also encouraged teaching to the test and narrowed what many schools taught. The gaps it sought to eliminate largely remained.
What AI can actually do
AI's advocates are right about some of its potential. Generative AI can help explain difficult concepts, provide immediate feedback, translate instructional materials, and make individualized support more accessible than ever before. Teachers can use it to differentiate instruction and reduce routine administrative work. Used well, AI will almost certainly improve teaching and learning in many classrooms. For educators looking to build these skills, AI for Teachers offers a structured learning path.
But what occurs in classrooms has never been the primary obstacle to educational equality. Students do not arrive at school with equal access to stable housing, nutritious food, quality health care, reliable internet, experienced teachers, safe neighborhoods, or family resources. These inequalities accumulate long before a child enters kindergarten and continue after the school day ends. Chatbots, no matter how well-designed, will not erase them.
The historian is not arguing against AI, though the drawbacks deserve attention - such as interfering with students' ability to think critically, which is no small thing. Schools should experiment with new technologies that genuinely help students learn, and teachers have always adapted to new tools, from calculators to the internet. The problem arises when technological innovation is mistaken for social policy. For more on how AI is being applied in classrooms and its limits, see AI for Education.
Why this matters for educators
For educators, the practical takeaway is to use AI where it genuinely helps - explaining concepts, providing feedback, reducing administrative load - without expecting it to solve problems it can't. History suggests that educational reforms can make some schools better, some teachers more effective, and some students more successful. What they can't do is eliminate the economic and social disparities that shape children's lives before they ever enter a classroom.
"AI may transform how students learn," the historian writes. "It will not, by itself, transform the unequal society that shapes students' lives and in which they learn."
If we really want to narrow opportunity gaps, we must stop expecting educational innovations - whether charter schools, standardized testing, or artificial intelligence - to accomplish what only broader economic and social reforms can achieve.
Your membership also unlocks:
