Skip to content

Are School AI Bans in LA and NYC the Beginning of a Backlash Against Classroom AI?

“No ban is going to keep AI out of students’ hands.”

Joseph South, Chief Innovation Officer at the Association for Supervision and Curriculum Development

Some of the country’s largest school systems are tightening the rules around generative AI. New York City Public Schools barred student-facing generative AI through eighth grade and limited high school use to vetted programs and AI literacy instruction. Los Angeles Unified placed a moratorium on student access through district devices while it reviews instructional uses and safeguards.

These decisions come after a period of rapid experimentation in classrooms, often ahead of clear guidance on what students should be allowed to hand over to AI. Questions about critical thinking, privacy, and assessment are now harder for districts to sidestep, especially when teachers may have little visibility into how much of a student’s work reflects their own understanding. Access adds another complication. Students can still use AI outside school, and the tools themselves are increasingly built into software they already rely on.

To better understand what is behind this shift, we spoke with Joseph South, Chief Innovation Officer at ISTE+ASCD and former director of the U.S. Department of Education’s Office of Educational Technology under the Obama administration. His work across education technology, policy, and classroom practice offers a useful view into why districts are drawing firmer boundaries around AI and what those choices could mean for students.

Meet the Expert: Joseph South, Chief Innovation Officer at the Association for Supervision and Curriculum Development

Joseph South

Joseph South, chief innovation officer, is a strategic national educational technology leader focused on evidence-based learning transformation.

He formerly served as the director of the Office of Educational Technology at the U.S. Department of Education. In his role at the department, he was an adviser to the Secretary of Education and developed national educational technology policy, formed public-private partnerships to assist state and local education leaders in transitioning to digital learning, helped school districts expand the use of openly licensed educational resources (OERs), and collaborated with stakeholders to nurture a robust ecosystem of edtech entrepreneurs and innovators.

He also worked on a cross-governmental team to bring high-speed broadband, interactive devices, professional development for educators and leaders, and high-quality affordable digital content to U.S. classrooms. He is a strong proponent of learners’ active use of technology.

South has led learning product development teams at startups, museums, nonprofits, corporations, and higher education institutions. He has also directed a host of learning programs and consulted on projects in China, Korea, Mexico, South America, and the Middle East.

He holds a doctorate in instructional psychology and technology from Brigham Young University.

Why Are Districts Hitting the Breaks?

For many districts, the immediate problem isn’t whether AI exists in students’ lives. It clearly already does. The harder part is deciding what schools should permit, what teachers should supervise, and where the line sits between useful assistance and work a student still needs to do independently.

“AI is a powerful technology, and it isn’t surprising that people have concerns about how it might be used in schools,” South says. “I am all for a thoughtful approach that considers both the risks and the benefits.”

South says the recent pullback in the United States’ two largest school districts- among others- reflects that uncertainty. School leaders face pressure to set rules around a technology that changes faster than most district policies do, and many do not feel ready to manage broad student use.

“I think what we are seeing is education leaders feeling unprepared to manage its use, so they are putting the brakes on,” he says. “It isn’t bad to pause a rollout to get your wits about you, but a long-term ban isn’t going to serve students well. A year is longer than I would recommend for a pause.”

And the limits of this approach show up quickly outside school, as blocking AI on district devices doesn’t stop students from using it at home, on personal devices, or through other software. That means the students most likely to keep experimenting with AI may also be the ones who already have better access to technology.

“Students need to learn how to master AI in support of their learning and growth if they are going to be prepared for the jobs of today and tomorrow,” South says. “And by banning it, you are just leaving students to figure that out on their own without any guidance from educators, because those that have means and access are absolutely going to keep using it outside of school hours.”

That leaves districts with a narrower and more difficult task. They still have to decide what good use looks like in the classroom, especially when the same tool can help a student move forward or do too much of the thinking for them.

Schools Still Need a Clear Line Between Help and Harm

Deciding what good AI use looks like becomes harder once the question moves from district policy into the work students actually do. For example, the same tool can help a student get past a minor obstacle, strengthen how they communicate an idea, or complete the intellectual work an assignment is meant to develop. In South’s mind, the useful dividing line is how much thinking the student still has to do.

“Learning requires productive struggle. We know this principle from getting proficient at sports or playing a musical instrument. You don’t achieve any kind of mastery unless you put in some real, focused effort,” South says. “Similarly, our brains need to be exercised to learn and retain and make long-term connections with subjects at school.”

Some of the difficulty students encounter has little to do with the skill or knowledge being assessed. 

As an example, South points out, “AI is useful when it supports [things like citation] by reducing the unproductive struggle that gets in the way of learning. Things like formatting sources in APA style take forever and don’t lead to meaningful learning.”

And AI can also give students tools to express an idea they already understand but may not have the technical ability to execute, like a student who wants to make an animation for a statistics report but does not know how to build one independently.

“We also shouldn’t overlook how AI might enhance storytelling and communication. Imagine a student who wants to add an animation to make a point more clearly in their statistics report. They may not have that skill on their own, but using AI, they are able to do it,” South says. “Being able to level up your report is exactly what an employer is looking for in a new hire and doesn’t compromise a student’s cognition. In fact, learning to use AI to create the animation actually enhances it.”

The bigger issue, it seems, is critical thinking and replacing actual learning. If the student hands over the reasoning or subject knowledge that the assignment is supposed to exercise. This arguably detracts from the students’ ability to learn something new and worthwhile.

“If a student is handing off essential tasks to AI, basically using it as a super efficient answer machine, then they don’t get the mental workout that leads to real learning. That is a terrible way to use AI,” South says. “Students who do this regret it later when they can’t perform on a test or when they realize that the content they didn’t learn is a prerequisite for a class or an experience that they really wish they had prepared for. It will catch up to you sooner or later.”

South also sees students use AI after they have already done some of that work themselves. “On a positive note, I have seen students who will use AI to prep for a test by typing out the answers they plan to give on the test and asking the AI for feedback,” he says. “Students also can use AI to get unstuck. For example, when they have tried really hard to figure out a problem and are just spinning their wheels and then ask AI for a hint or to help them see where they went wrong in their thinking.”

Even so, the range of uses makes a blanket rule difficult to apply across subjects and assignments, with so many varied uses and applications of AI in learning and output of demonstrated or performed learning. Schools first have to know which parts of the work students need to own themselves, and where AI can enter without taking that work away.

Bans May Be Easier to Announce Than to Sustain

Drawing a boundary around AI in the classroom is one thing. But keeping that boundary intact once students leave school is much harder. Districts control their own devices, networks and approved software, but students move between school technology and personal phones, laptops and accounts throughout the day.

South is skeptical that access restrictions can reach much further. 

“No ban is going to keep AI out of students’ hands. Students with means and access will absolutely keep using AI outside of school hours,” he argues. “But students who don’t have the means or access will not.”

Students who can reach AI elsewhere continue learning how to use it, often without much direction. Students who depend on school technology lose the opportunity to build the same familiarity. They also lose access to teachers who could help them understand when the technology is useful and when it begins doing too much of the work.

“Ultimately they will be at a disadvantage because they won’t be getting the experiences and guidance to help them use AI in productive ways to enhance their learning and growth that the school could have been providing,” South explains. “If only those with means and access have the opportunity to learn to use AI, we are going to greatly expand the opportunity gap.”

AI is also becoming part of the skills students may be expected to do later in school and at work. The more experience some students get with those tools now, the more familiar they become with where AI can help and how to use it.

“The bottom line is that every future job will require AI use and those with experience using it for things like research and communication and storytelling are going to have a huge head start over those who don’t have the opportunity for that,” he adds. “Schools should be creating those opportunities and helping students use AI wisely rather than banning it altogether.”

Districts can decide what students access on school property. They have much less control over who gets to practice with AI everywhere else. This makes the next step harder than simply lifting a restriction: schools still need rules for how students use AI once it is allowed back into the classroom.

Schools Need Rules That Can Change With the Technology

Moving forward, districts need rules that distinguish between assignments where AI can support the work and those where students need to work without it. Those expectations can also be set at the classroom level, where teachers and students have a clearer sense of what an assignment is meant to develop.

South recommends five main areas for schools to focus on: clear guidelines, transparency, AI literacy, teacher training, and changes to assessment.

“First, we need clear guidelines for when and how to use AI and when not to. And students should be part of those conversations and help create the norms. That way they will be more invested in following them,” South explains. “Also, not every assignment needs to allow AI at all or to the same degree as another assignment. This can be clearly communicated assignment by assignment, so students and the teacher are on the same page. This ensures that productive struggle will continue to be a part of learning, but also opens the door for new, powerful uses of AI to support learning.”

A workable policy also depends on disclosure. If students and teachers are open about where AI enters their work, schools have more information to draw on as classroom practices change.

“Second, we need a culture of transparency of AI use. That means that both the educators and students disclose how and when they use AI in their work. This creates mutual accountability as well as a paper trail that can be used to evolve and adjust the policy over time.”

AI literacy can fit within the digital citizenship instruction already taking place in schools. “Third, we need to incorporate AI literacy into our digital citizenship efforts we already have underway,” South continues. “Responsible use of technology should already be a regular part of the conversation at school, and AI is another powerful digital tool that requires similar discipline and awareness to use effectively.”

Teacher preparation is another part of making those policies work in practice. “Fourth, we have to train our teachers. They need to feel confident using and talking about AI with their students. If they don’t feel prepared, then none of the other steps are going to ultimately be successful.”

Assessment may require the larger change. Essays, reports, and other finished assignments let teachers judge what students understand, but generative AI makes the final product less revealing when similar work can be produced without the same learning process behind it.

“Fifth, we need to evolve our assessments. Grading a final product was always a proxy for student learning. Now the final product can be easily mimicked by AI,” South notes. “We need to evolve our assessments to focus more on measuring and improving the processes that get a student to a final product. Those are harder to simulate with AI and are where most of the real learning occurs anyway.”

The current restrictions give districts room to work through some of these questions, but access is only one piece of the policy schools now have to build. How AI is used in assignments, what students learn about it, how prepared teachers are to guide them, and what counts as evidence of learning will determine what classroom AI looks like after the bans.

Chelsea Toczauer

Chelsea Toczauer is a journalist with experience managing publications at several global universities and companies related to higher education, logistics, and trade. She holds two BAs in international relations and asian languages and cultures from the University of Southern California, as well as a double accredited US-Chinese MA in international studies from the Johns Hopkins University-Nanjing University joint degree program. Toczauer speaks Mandarin and Russian.