Colorado students call for banning generative AI at their high school: grades, water and art at the heart of the debate

By Zendoric · October 8, 2026. Two students at Montezuma-Cortez High School (MCHS), in Cortez (Colorado), asked the Re-1 school district board on Tuesday to ban the use of generative artificial intelligence (AI) at the school. Peyton Perry and Will Blair presented the results of their own survey of 49 students and five teachers, as community radio station KSJD reported. We are not talking about a technology board or a regulator: these are two teenagers asking to be protected from a tool that, they said, has already spread "exponentially" through their high school over the past year, to the point that its use has sometimes been required in classwork.

The first argument was the cognitive one. Perry cited a study she attributed to the American Psychological Association (APA, the US psychologists' association), according to which minors who use AI frequently show worse exam performance, lower brain activity and a reduced ability to generate new neural connections. The school has spent several years trying to raise its exam results, the student recalled: if the diagnosis is correct, the school's stated goal runs head-on into the tool it is adopting. Zendoric has not been able to verify the specific study Perry cited, but the underlying concern is not marginal: in general, there is already a body of research -from the MIT Media Lab to several universities- on so-called "cognitive offloading," the tendency to stop exercising one's own memory and reasoning when a machine solves the problem for us.

The second front was water. Perry stressed that her region, in the southwestern United States, is "highly stressed" in water resources, and that the data centers that train and run these models consume water directly -for cooling- and indirectly, by driving up electricity demand. She cited a study according to which, if all the data centers currently proposed in the Phoenix area are built, consumption would go from 385 million gallons a year to 3.7 billion, and some of that water could come from tributaries of the Colorado River, among them the Dolores, which runs through her own county.

The third argument, made by Blair, was intellectual property. In his view, school staff using generative AI to create posters and materials violates the school's own copyright policy because, in his words, "what AI does is steal" and the way it accesses other people's art is in itself unethical. He cited as an example that companies such as Midjourney face lawsuits -still unresolved in US courts- accusing them of training their systems on protected works without authorization, and he asked that the school assign that work to its own student artists instead of to an image generator.

The teachers interviewed qualified the picture. According to Perry, they acknowledge that AI can be useful for analyzing data, but believe it limits their ability to connect individually with each student. And some admitted something more uncomfortable: with a workload that has risen "without time or compensation rising," they feel that using AI is by now almost mandatory, "even if morally they are against it." This is the pincer that defines much of AI adoption in the public sector: it is not always chosen by those who end up using it.

What is striking about the Cortez case is not the content of the arguments -the debate over cognition, water and copyright has been circulating in the education sector for some time- but who is putting them on the table. It is not the parents, nor a teachers' union, nor a state regulator: it is two teenagers, the generation that, according to the narrative AI companies themselves tend to sell, should be the most enthusiastic about these tools. When skepticism comes from within the group supposedly set to benefit most, it deserves more attention than when it is repeated by an adult with already-formed prejudices.

At Zendoric we argue that, in the long run, AI can free up time and generate abundance, in education too, if it helps personalize learning and frees teachers from administrative tasks. But this episode is a good reminder that the promise does not fulfill itself: it is fulfilled, or frustrated, depending on how it is implemented. A district that goes as far as requiring the use of generative AI in classwork without first investing in training, without easing the teaching workload and without resolving the intellectual property debate is not distributing abundance: it is shifting the cost of the transition onto those with the least power to negotiate it, whether students or teachers.

The Re-1 board limited itself to thanking them for the presentation, without announcing any policy change. But the pattern Perry and Blair describe -accelerated adoption, with no clear framework, imposed rather than agreed- is repeating itself in districts across the country. More than the outright ban these two students are asking for, that is what really ought to be watched: not whether AI enters the classroom, which it already has, but whether it enters with the safeguards, the training and the time that both students and teachers say they lack.