Educators warn classroom AI could deepen concerns over school screen use

Generative AI is becoming common in schoolwork, and some education researchers say schools have not put enough safeguards around it. Their concern, according to Fortune, is that chatbots may repeat problems seen during earlier waves of classroom technology by letting students bypass the mental work needed to learn.

More than 80% of high school students said they used generative AI for school assignments last year, according to a College Board survey cited by Fortune. The nonprofit, which runs the SAT and AP programs, collected the finding after ChatGPT’s 2022 release helped bring AI tools into everyday student use.

Researchers raise concerns about learning

A Brookings report published in January reviewed interviews and focus groups with more than 500 educators, parents and students in 50 countries, along with more than 400 studies, Fortune reported. The researchers concluded that, at this stage, the dangers of using generative AI in children’s education outweigh the benefits.

The Brookings report pointed to early research on AI and cognition, including a February 2025 Microsoft study that found AI use was linked with weaker judgment and critical thinking, according to Fortune. Mary Burns, an education consultant and co-author of the Brookings study, told Fortune she believes AI-related cognitive offloading can contribute to declines in critical thinking, reading, writing and basic factual knowledge.

The debate comes as broader school technology use faces new criticism. Neuroscientist Jared Cooney Horvath told Congress in January that data from the Program for International Student Assessment show Gen Z is the first modern generation to test as less cognitively capable than its parents, Fortune reported.

Horvath attributed the shift to broad access to classroom technology and said lower standardized test results correlate with more screen time in school, according to Fortune. A 2014 study of 3,000 university students found that students spent about two-thirds of screen time on activities unrelated to class.

Earlier edtech efforts are under review

Horvath told Fortune that the modern school technology push gained strength after Maine started a statewide student laptop program in 2002. Google later worked with educators to test inexpensive Chromebooks with free apps and encouraged school staff to promote them, according to Fortune; by 2017, Chromebooks made up more than half of digital devices shipped to schools.

Horvath, author of The Digital Delusion, also pointed to a longer history of automated instruction. He cited the 1924 “teaching machine” invented by Ohio State University psychology professor Sidney Pressey, saying students could perform well with the device but struggled to apply the knowledge away from it, Fortune reported.

He made a similar argument about AI: tools that help experts work faster may not help beginners build skill. Horvath told Fortune that students who rely on AI without clear limits may learn dependency instead of the underlying subject.

A University of Wisconsin survey published last month found that fewer than one-third of 303 educators and school professionals in Wisconsin and 132 education professionals elsewhere in the United States had put AI policies or guardrails in place, according to Fortune.

Some educators see limited uses

Burns, who supports some education technology, told Fortune that banning AI outright is not realistic. The Brookings report found that teachers are using AI to prepare lesson plans even as they worry students will use it to cheat.

Burns said classroom AI data remains limited but added that the tools can help in specific cases, such as allowing teachers to adjust the reading level of a passage for English language learners. Her view, as reported by Fortune, is that technology in schools has produced mixed results rather than a clear success or failure.

This story draws on original reporting from Fortune.