3 Tips to Cover the Risks of AI in Education

New York City, home to the nation’s largest public school district with nearly 1 million students, banned artificial intelligence usage for students in preschool through the eighth grade. 

The New York City Department of Education’s new policy, announced on Sept. 2 for the 2026-2027 school year, also recommends strict screen time limits for third to fifth graders. 

The ban covers generative AI as well as “chatbots” and AI tutors. “Younger students need hands-on learning, social interaction, play, and teacher-guided instruction,” according to the New York City Education Department policy documents obtained by Chalkbeat. “Routine individual screen use can crowd out those experiences.”

Certainly, the AI landscape is changing rapidly in both K-12 and higher education. For education reporters looking to stay on top of these issues, here are three things to keep in mind:

1. Pay Attention to the Research. 

Studies have long documented the effects of “cognitive offloading” – delegating decision-making, memory, and information retrieval to outside systems. Humans are essentially giving up the brain development that comes with keeping those tasks “in-house.” Some researchers contend the risks outweigh any potential rewards, particularly for younger students, that can come with AI use. 

Separate from the longer-term questions about brain development, the findings are mixed on how effective AI is at improving student learning in both K-12 classrooms and higher education. 

A new study of Chinese students in grades seventh through 12th found they significantly improved their homework productivity using AI, but then saw their in-class test scores crater. 

Reporters don’t have to look overseas for anecdotal evidence to support the findings. A Brown University economics professor made his usual midterm exam a take-home test for the first time this past academic year, and he quickly regretted it. After a higher-than-usual percentage of students had perfect or near-perfect scores, he suspected AI usage, in violation of the university’s policies. He switched the final exam to a written test in the classroom – and more than a dozen students promptly withdrew from the course, despite their “A” standing, according to Inside Higher Ed.

Dana Goldstein of The New York Times has a thoughtful piece looking at what students really lose when they stop writing independently. The surge in college professors returning to handwritten bluebook exams to deter AI assistance on take-home tests is “baffling to tech industry leaders, who see A.I. rejection as a form of educational malpractice,” Goldstein wrote. 

This is an opportunity: Watch what they do – and not just what they say. Silicon Valley executives apply very different rules for their children’s learning environments than what their companies push to younger consumers. 

From The New York Times’ David Streitfeld:

Phones are banned at the Waldorf School of the Peninsula, two miles from Google’s headquarters. “No phones means no distraction, no social media anxiety, and no constant comparison. Just real childhood,” the school explains on its website … Several Waldorf board members are professionally focused on artificial intelligence, but at the school[,] machine learning is minimized in favor of ‘human learning.’ The goal is to train the future elite to think clearly.

The debate isn’t just about how students interact with the technology. Recent studies have found AI is not a substitute for a human teacher, either for daily classroom instruction, lesson planning, or tutoring. Even so, some teachers are going “all in” on generative AI to help them manage their workloads. That’s raising concerns about AI introducing errors into teaching materials educators buy from AI services online.

Does your school district offer a virtual learning program or digital high school? Has a private Alpha School campus popped up in your neighborhood? For public schools, do an open records request for student demographic data (stripped of any identifying personal details that might violate the Family Educational Rights and Privacy Act, or FERPA). Ask to shadow a student for a week, and see what those environments are really like. (If the schools turn you down, that’s worth reporting, too.)

2. Universal AI in Education Is Not a Foregone Conclusion.  

New York City is far from alone in restricting AI student use. (Entire countries have taken aggressive steps, including Norway.) 

In Los Angeles, where the nation’s second-largest school system resides, leaders have banned the use of AI classroom tools by students younger than 13. Older students are eligible to use AI, only with a teacher’s permission, and after they complete a lesson on “digital citizenship,” as the LAist reported. 

The Los Angeles Unified School District formed a committee to explore stricter limits for AI use in other educational settings, and it is soliciting public input through a series of community meetings. What’s the current policy for AI usage in your school district? How does it differ by grade level? Are school board candidates campaigning on these issues for the midterm elections in November? In Chicago, a parent group is asking school board candidates to sign a pledge that they’ll back strict, three-year limits on AI in classrooms. (Not everyone is on board.)

At the other end of the academic spectrum in the Windy City, the University of Chicago recently banned AI in its core social sciences curriculum. As The Chronicle of Higher Education’s Shea Vance wrote, “The new policy is a welcome change for faculty members who have grown frustrated with AI’s looming presence in the classroom. Aaron Jakes, an associate professor of history who teaches a core social-sciences course each winter, said his initial reaction to the memo was ‘relief and jubilation.’”

While some U-Chicago faculty might be relieved and jubilant, their peers at many other campuses say there’s still too little guidance for when, and how, to allow AI use by students, or to incorporate it into their own work. (See The New Yorker’s “The Despair of the Professor in the Age of A.I,” and this roundup from EdSource with perspectives from California.)

Who’s in charge of campuswide AI policy at the colleges and universities in your community? Are public town halls and faculty forums scheduled to discuss looming changes, and when are they taking place? 

Who are the largest employers in your community, and what are their policies for AI use in their workplaces? How do those expectations align with, or differ from, the rules students are being told to follow until graduation? How are postsecondary educators, in particular, looking to address that disconnect?

3. Talk With Students About What Drives Them to Use AI. 

Greg Toppo at The 74 has a fascinating roundtable with high school student journalists: They say AI – and related fears about plagiarism and cheating – is changing the culture of their high school newsrooms. With support from teachers, students are crafting policies that set parameters for acceptable AI usage (to search for background on sources or to spell check, for example) rather than blanket bans.

In Boston, students from all 50 states gathered to craft AI rules they hoped public schools would adopt. The lack of a national policy on AI leaves the door open to leeway – and confusion, as NPR’s Cory Turner reported. “We missed the mark with social media and phones,” Hunter Wurzel, a high school senior representing Ohio, told Turner. “That’s been a huge talking point already [this weekend], is how we failed to implement those things into education and teach the youth how to use them.”

Need a jumping-off point for your interviews with students? Look at Inside Higher Ed’s recent survey of college students on AI usage. Many students have mixed feelings about generative AI’s effectiveness and impact on their learning, even as a growing percentage say they’re using generative AI on a weekly, if not daily, basis. (The percentages in the IHE survey are higher for students in four-year colleges than community colleges, an interesting data point worth localizing.) 

Some young people view AI as a practical shortcut to “busywork” that has little purpose or meaning. Others believe they will fall behind their peers who are already using it. Many young people fear AI will replace them in the workforce. 

An August Pew Research survey found, for the first time, that more than half – 55% – of young Americans under 30 were “concerned” about AI, rather than “excited” by its prospects. Separately, many younger people – and people of all ages – oppose AI for environmental reasons. Those concerns should be factors in your reporting, as well.