Big-district AI restrictions shouldn’t overshadow opportunities to save teachers time, strengthen instruction and connect with students, experts say.
Current headlines may suggest that education is still struggling to define AI’s role in the classroom.
While Los Angeles USD and New York City Public Schools have limited students’ access to generative AI, both districts allow educators to use the technology for lesson planning, record-keeping and other purposes.
That distinction matters. Students must learn to use the technology that’s reshaping school, work and daily life.
AI literacy informs students about AI’s limitations and biases. It can also reinforce the durable skills classrooms have taught for decades, says Chris Agnew, director of the SCALE Initiative’s AI for Education Hub at Stanford University.
“Those skills have been escalating in value over the last 20 years with the increasingly digital world, and AI is only fueling that more,” Agnew explains.
Restricting access without building AI literacy also risks leaving underserved students unprepared for an AI-driven job market, argued Janice Jackson, director of the Aspen Institute’s Education and Society Program, in a recent episode of the “DA Changemakers” podcast.
“Leaders should definitely look at screen time usage based on a child’s age and their developmental state,” she said. “However, I don’t think you should totally ban it. I think that young people can learn to use technology early.”
An ‘adults-first’ AI strategy
Teachers who use AI to handle administrative and routine tasks can spend more time with students.
A new report from the SCALE Initiative found that nearly 87,000 of MagicSchool’s most active users primarily relied on AI for routine tasks, including communication, lesson planning, feedback and administrative work. The teachers analyzed in the study were among the platform’s top 5% of users, having used it on at least 52 days between May 2023 and November 2025.
The study also found that teachers prefer tools designed for specific classroom workflows over general-purpose chatbots like ChatGPT.
An educator-first approach can enhance instruction while leaders set boundaries around student AI use. In Washington’s Peninsula School District, teachers are offloading administrative work to personal AI agents to “spend more time being human,” said CIO Kris Hagel.
One of the district’s high school chemistry teachers used AI to simulate cell membranes and build interactive learning games. Hagel sees room for AI across the curriculum, from CTE to financial literacy.
“We’re fascinated by what we can build, and we’ve got a really strong foundation here,” Hagel says. “Now, it’s just getting it in front of people so they can use their creativity to see what they can come up with.”
Eliminating ‘shadow AI’
As AI adoption grows, leaders must also guard against “shadow AI,” the ungoverned use of platforms by students and staff, says Tucker Austin, K-12 AI market and product strategist at SchoolAI.
The risks are legal as much as pedagogical: If a teacher uploads a student’s IEP into a personal chatbot, that platform can train on the data, violating FERPA and COPPA, Agnew adds.
Austin recommends districts “look at vendors and providers with a very scrutinous eye” and select providers with zero data retention policies, requiring vendors to discard users’ input prompts and their responses.
“Teachers are adopting free AI products all over the place,” he says. “The number one thing that we recommend districts do is select a provider that can give them a general-purpose AI that can support teachers as well as create experiences for students.”
For district leaders, the challenge may not be deciding whether AI belongs in schools, but determining where it can provide value, where students need guardrails and how educators can use it responsibly.
