A school choice revolution has swept America. Following pandemic-era school closures and culture wars, choice proponents have made unprecedented gains at the state and federal levels. Today, 34 states have adopted some form of school choice, with 1.5 million students participating.
The rise of widely available artificial intelligence (AI) tools constitutes a quieter but no less profound education revolution. Large language models (LLMs) like ChatGPT can deliver quality, individualized education at little to no cost. Teachers increasingly use the technology to create lessons, assignments, and assessments, and educational technology firms are rapidly integrating AI into their products.
Finally, a reason to check your email.
Sign up for our free newsletter today.
As AI advances, the cost of pure instruction—unbundled from childcare, mentoring, and other human services—will fall dramatically. AI tutors like Khan Academy’s Khanmigo and SpaceX’s Synthesis cost just $4 and $29 per month, respectively. Tools like these will empower leaner models of education—such as micro-, home-, and hybrid-schooling—and help build out the growing school-choice movement.
Alpha School, an innovative private school with locations across the country, offers a glimpse of where this policy-tech combination may be headed. Alpha students spend two hours every day studying a digital, AI-integrated curriculum. They spend the rest of their time developing life skills and pursuing passion projects. The adults in the building, called “guides,” don’t teach; instead, they provide oversight, leaving the subject-matter instruction to the technology and, in some cases, on-call tutors.
Most of Alpha School’s current offerings are expensive, but the company is growing and diversifying. Today, it offers micro-campuses for student-athletes and is planning low-cost schools that will be covered almost entirely by state vouchers. Other institutions, like Primer Microschools, are proving it’s possible to offer a similar model of micro-schooling to middle- and low-income families in states with education savings accounts.
As more schools embrace individualized and AI-powered instruction, students will be free to move through the curriculum at their own pace, allowing them to master the material. This development will also make grade-acceleration policies—which allow students to advance to the next grade immediately after completing their current grade’s coursework—more common.
Of course, we should be wary of AI’s effects. For one, not all AI tools are created equal. Some tools, like learning platforms with generative AI integrated, give feedback after students answer a question. A standard LLM, by contrast, can simply give students the answer outright—facilitating cheating, and undermining kids’ ability to learn the material.
We also haven’t seen AI adopted at scale. Alpha School claims its students learn 2.6 times faster than average, but as the technology reaches more classrooms, we may find that the average student struggles to work with AI as he would with a human teacher. Some students may learn best from group activities such as call-and-response, which aren’t possible in the personalized, AI-driven approach.
School-choice proponents must also reckon with AI’s ideological biases. Popular LLMs lean left, and as the technology reaches more classrooms, companies, regulators, and legislators will be tempted to entrench that bias. These models’ political biases could undermine educators’ ability to promulgate their school’s values.
The solution to these issues is markets. Much as parents have had to educate themselves on phonics, critical race theory, and other pedagogical topics to evaluate what their children are learning in school, parents will also have to learn about AI when choosing a school for their children. If parents demand it, the market will provide AI tools and programming tuned to various value systems: liberal and conservative; secular and theistic; strict and supportive.
Homeschoolers, micro-schoolers, and others on the education frontiers should experiment with AI-powered models. Education researchers should study those models’ effectiveness, and policymakers should work to ensure a competitive, pluralistic playing field. Now is the time to test AI’s potential in education—and to build the future of the school-choice movement.
Photo by Courtney Perry/For the Washington Post