Column
AI instruction should be harm-reduction through educating students about the pitfalls.
By Eric Stinton
June 21, 2026 · 7 min read
Eric Stinton
Eric Stinton is a writer and teacher from Kailua. You can follow his work through his newsletter at ericstinton.substack.com.
AI instruction should be harm-reduction through educating students about the pitfalls.
Artificial intelligence is here to stay. Or so I’m told.
I’m skeptical, especially since the sentiment seems to have originated from AI companies themselves. Take a quick look around society, and you don’t find that kind of certainty or consensus on the ground.
Most voters say the risks of AI outweigh the benefits. Gen Z has become less excited and more angry about these tools. College graduates are booing commencement speakers for mentioning AI in a positive light. A strong majority of Americans oppose building AI data centers in their communities.
Republicans and Democrats alike are overwhelmingly in support of greater regulation, even if it slows innovation. The Pope sees the AI industry as the “idolatry of profit that sacrifices the weak,” while tech CEOs – yes, tech CEOs – see the use of AI as a sign of laziness and dishonesty.
Just think about how you encounter AI in your life. If someone says “check out this picture, it looks like AI,” are you expecting it to look good? If you get an email that you suspect was written by AI, do you feel awed by the technology, or insulted that you have to spend time reading something that someone spent no time writing?
But apparently AI is inevitable! Which is why most AI services are forced upon users – who, anecdotally, are almost always annoyed by it – and voluntary usage of LLMs drops off dramatically during summer, when students are no longer using it to cheat … err, no longer using it as a tool to assist them with their work.
I do not have enough space here to get into AI-assisted suicides, severe environmental degradation associated with AI technology, or the fact that AI companies are hemorrhaging money, floating on venture capital and government subsidies until they can go public and build enough exit liquidity for early investors to cash out and bail. To say nothing of the absolutely insane, dystopian things AI execs are publicly saying on a regular basis that should give everyone pause.
Still, there are a number of ways to interpret the phrase “AI is here to stay.” I don’t necessarily disagree with the sentiment, but not because I’m sold on the technology so much as I’m resigned to the fact that the wealthy and powerful in this country tend to get what they want, regardless of how most people feel.
It was nonetheless interesting to read about how Hawaii schools have been adapting to AI. It’s a disruptive technology, and it’s important that we have open discussions about best practices as we collectively learn and adjust to its prevalence. While I appreciate the openness and honesty, much of what I saw makes me anxious.
Winston Sakurai, educational administrative services director in the state Department of Education’s Office of Curriculum and Instructional Design, said “we want people to embrace it, but we cannot afford to burn people out with the adoption of it.”
On the surface this sounds reasonable, and I don’t want to look too much into a short quote, but it also sounds like there’s an expectation for teachers to adopt it eventually. This is a big red flag, not just because it offers relatively frivolous benefits – image generators! Chatbots! – in exchange for a mountain of harm across a variety of measures. My feeling, too, is that once teachers are expected to outsource some of their workload to AI, more work will be shoveled onto them, even as AI continues to require human input and cross-checking – aka work.
Meanwhile Mid-Pac is using it in preschools, in a teacher-controlled setting, it’s worth noting, to point out that it makes mistakes and is “not the gospel, but a tool,” as the teacher said. I applaud the thought, but I fear – and teachers across the country do, too – the lesson that will stick is not the nuanced one that AI is a flawed tool, but that AI is the first-stop to make for any inquiry. A thought that surely warms the hearts of the most predatory tech execs around the country.
The pattern I see most with the high school and college students I teach is that they know AI makes things up, but it’s a risk they’re willing to take to avoid doing work they’d rather not do. If the promise of AI is to remove tedious unnecessary work, eventually most work becomes seen as tedious and unnecessary. I’m not convinced early exposure to AI tools will do anything but solidify that rationale.
Andy Gokce, executive director of Kūlia Academy, said “we want (students) to know, inside and out, how it actually works.” That’s a great idea, and I hope it includes the environmental and economic damage caused by the AI industry and an analysis of the ethics of a technology built on scraping data from other people to enrich a select few.
I would also hope such instruction includes research that is coming out that shows how students who use LLMs to assist them with their essays, as Gokce highlighted, “consistently underperformed at neural, linguistic, and behavioral levels,” and students who use AI to help them study “scored significantly lower” than those who do not use it.
They should investigate the role that AI usage plays in the unfolding crisis of university students across the country who struggle with middle school math and basic reading. They should research how using AI to offload work ends up removing the very processes that actually build cognitive capacity and concentrative stamina.
As more data emerges that will likely provide more evidence for these kinds of findings, I hope schools are responsive, though inconvenient data tends to get ignored, no matter how often schools say they’re data driven.
There’s a fear that if students don’t get used to the technology now, they’ll fall behind. This is nonsense. LLMs are designed to be as easy to use as possible; there’s no technical bar to clear. If there is a skill involved, it would be how you prompt it. Learning what kinds of questions to ask and how to ask them has been standard issue education for as long as I’ve been alive; we don’t need to infect that skill acquisition with AI.
I’m not suggesting that schools should abandon teaching students how to use AI. But instead of thinking of AI as a vital technology, the way learning how to use a computer was for my generation, there’s a better, more accurate model: harm-reduction.
I liken AI instruction to the efforts of needle exchange programs, where people worked with drug users to educate them on the harms of drug use, as well as provide clean needles to reduce the spread of infections. There is a built-in understanding that AI use, like drug use, exists in our world, even though it’s on the whole bad for people and bad for society.
Measures can be taken to reduce the associated harm, but ultimately the best possible outcome is not using at all.
Civil Beat’s education reporting is supported by a grant from Chamberlin Family Philanthropy.
· June 21, 2026 · 1 min read
Support timely, accurate, independent journalism.
Honolulu Civil Beat is a nonprofit organization, and your donation helps us produce local reporting that serves all of Hawaii.
Eric Stinton
Eric Stinton is a writer and teacher from Kailua. You can follow his work through his newsletter at ericstinton.substack.com.
Read more
Ideas is the place you’ll find essays, analysis and opinion on public affairs in Hawaiʻi. We want to showcase smart ideas about the future of Hawaiʻi, from the state’s sharpest thinkers, to stretch our collective thinking about a problem or an issue. Email news@civilbeat.org to submit an idea.
Get the Civil Beat App.
Sign up for our FREE morning newsletter and face each day more informed.
Copyright © 2010-2026 Honolulu Civil Beat Inc. All rights reserved. Civil Beat ® is a registered trademark of Honolulu Civil Beat Inc.
You’re officially signed up for our daily newsletter, the Morning Beat. A confirmation email will arrive shortly.
In the meantime, we have other newsletters that you might enjoy. Check the boxes for emails you’d like to receive.
Inbox overcrowded? Don’t worry, you can unsubscribe
or update your preferences at any time.
Leave a Reply