Here is an archive link for those who cannot get around the paywall: https://archive.is/PwNH9
AI generated summary for those who, umm, appreciate the irony:
The author, whose kids attend public schools in Massachusetts, writes about the rapid and largely unconsented rollout of AI tools in K-8 classrooms — from Amazon-branded “AI certificates” given to third graders, to Google Gemini being pre-installed on school-issued Chromebooks and nudging sixth graders to let it write and design their assignments for them.
The piece lays out the growing research concerns: studies from MIT, Carnegie Mellon, and Oxford suggest that students who rely on AI for problem-solving perform significantly worse and give up more easily once the AI is taken away. Researchers also warn that chatbots, which mimic emotional intimacy, may interfere with the social development that is so critical in the tween years. The broader argument is that AI shortcuts the very struggle that builds thinking, creativity, and persistence in young people.
On the corporate side, the article raises serious questions about Google’s Gemini being made available to children of all ages, about school officials receiving fellowships funded by Google and ed-tech investors, and about teacher and parent voices being largely left out of the conversation.
Some parents and educators are pushing back through coalitions calling for a two-year pause on AI in K-12 classrooms. The article closes with a simple but pointed question: what if what we want from AI in our kids’ schools is simply nothing?
AI in your kids’ classroom — where do you stand?
- I’m fine with it — AI is the future and kids should learn it early
- I’m cautious — some uses are okay but it needs much stricter limits
- I’m opposed — kids should develop their own thinking first
- I don’t have kids in school but I have a strong opinion anyway
| Age Group | Currently Deployed | What Researchers Say |
|---|---|---|
| All ages (incl. K–2) | Google Gemini: Pre-installed on school Chromebooks with no minimum age. “Help me write” and “Beautify this slide” prompts appear in every assignment. | The CEO of AI for Education said making Gemini available to all ages “so clearly” risks harm to children. Researchers warn chatbots mimic emotional intimacy before children have developed social skills. (Bickerstaff / New Yorker, 2025) |
| Ages 5–10 (K–4) | Amira reading bot: Records children’s voices in real time to give AI-driven reading feedback. Used in NYC, LA, and other districts from kindergarten. | “Children should not be using chatbots under age 10. These tools require expertise and evaluation skills that even many adults don’t have.” (Amanda Bickerstaff, AI for Education) |
| Ages 5–10 (K–4) | Adobe Express for Education: AI image generation available in 2nd grade art classes. Generated sexualized images when 4th graders used it for a book cover project. | AI integration may cause “cognitive atrophy” — and effects on a brain still building those muscles are potentially far greater than on adults. (MIT, 2025) |
| Ages 10–12 (Grades 5–6) | ChatGPT / Claude: Used by 6th graders in Boston public schools to prepare for state standardized tests. | Ages 10–11 are a critical window for social development driven by surging oxytocin and dopamine. A fawning chatbot “hijacks the biological tendency to want peer feedback,” stunting skills used for the rest of life. (Mitch Prinstein, UNC Chapel Hill) |
| Ages 10–12 (Grades 5–6) | Gemini in Google Slides: Rewrites and “beautifies” student presentations automatically, including adding its own images and resetting typography. | Students who used LLMs for math and then lost access “perform significantly worse without AI and are more likely to give up.” Persistence is one of the strongest predictors of long-term learning. (MIT / CMU / UCLA / Oxford, 2025 (preprint)) |
| All K–8 | Broad rollout: ~80% of K–12 teachers report their districts use Chromebooks, creating a near-universal market for Gemini. No consistent parental consent process in most districts. | A Brookings Institution analysis of ~400 research studies concluded that AI tools “undermine children’s foundational development.” About 1 in 5 student AI interactions involved cheating, self-harm, bullying, or other harmful behaviors. (Brookings, 2025 / Education Week analysis) |
