How to Use AI Without Getting Dumber
Left 100%
Center 0%
Right 0%
3 left · 0 center · 0 right
What happened
The Wall Street Journal, citing interviews with more than 60 teachers, school officials, parents, children, and clinicians, reported that students have become reliant on AI tools including Google’s AI-assisted search and Gemini chatbot, which come preloaded on Chromebook laptops supplied to schools. Google provides educational software to more than 170 million students and teachers worldwide. An OECD survey of 760,000 15-year-olds in 91 countries found average student test scores at their lowest in 25 years, with U.S. scores down 14 points from 2022, and students who reported using AI tools daily scoring 20 points lower on science tests than those who reported never using AI. A Google spokeswoman said the company’s internal studies show AI tools are a useful supplement to teacher-led classrooms and that teachers use the technology because it “solves real classroom challenges and helps improve learning outcomes.”
Omitted — what each side leaves out
Unpacked
Left-leaning coverage never reports Google's classroom AI rollout or the international score drop that right-leaning coverage treats as the story. A reader is told that chatbots can quietly erode writing and memory if used as a crutch. That reader is not told Google already reaches more than 170 million students and teachers. The same coverage skips a global teen survey whose scores are at a 25-year low.
Right-leaning coverage leads with Google's school push and kids left "dependent" on chatbots. Left-leaning coverage leads with advice for using AI without getting dumber, and treats classroom results as one example among many.
Neither side asks how much of the score drop comes from Google's classroom tools rather than from ChatGPT and other chatbots.
Bottom line
Vox never mentions that Google already serves 170 million students and teachers. The New York Post never mentions that students who used a chatbot scored higher with it and worse once it was gone.
The receipts — 6 facts one side's coverage lacks
- Google provides educational software to more than 170 million students and teachers around the world.
- Students who admitted using AI tools daily for school scored 20 points lower on science tests than students who reported never using AI.
- A study of nearly 1,000 high school students found ChatGPT users scored 48 percent higher on a math test with the tool, then 17 percent lower than the control group when retested without it.
- A Google spokeswoman said the company's internal studies show AI tools are a useful supplement to teacher-led classrooms and help improve learning outcomes.
- Former Google research scientist Mikel Rodriguez said the company is doing an experiment on an entire generation's cognitive development with no control group.
- Google has Gemini for Students, which is designed to act like a tutor by asking questions and encouraging students to find solutions rather than simply providing answers.
The Left View
Vox frames the phenomenon as cognitive offloading and “de-skilling,” not as a Google-specific school failure, arguing that chatbots deliver a short-term productivity lift followed by a “hangover” in memory, writing, and independent problem-solving. It cites a study of nearly 1,000 high school students in which ChatGPT users scored 48 percent higher on a math test with the bot but 17 percent lower than a never-exposed control group when retested without it, and University of California Irvine research finding students did better on text-based questions after ChatGPT’s arrival but worse on later offline tests. Experts in the piece say learned skills atrophy without practice and that users may not notice because AI makes them “believe they’re more capable than they really are”; a patent-lawyer working paper and a Middlebury experiment are used to distinguish “augmentation” users, who later outperform peers, from “automation” users, who do not. Vox’s through-line is that “ChatGPT brain isn’t inevitable”: the long-term solution “isn’t necessarily less AI” but using it intentionally—for tasks one can already do, in tutor or study modes, with friction and metacognition—while a Gallup/University of Pennsylvania survey found 68 percent of U.S. Gen Z adults worry that offloading cognitive tasks is “robbing them of important skill-building.”
The Right View
The New York Post, amplifying the Journal’s reporting, frames Google’s “aggressive AI push in schools” as having “caused many students to become ‘academically and emotionally dependent’” on chatbots and as contributing to a plunge in grades. It quotes Pennsylvania eighth-grader Richard Scott saying AI “slowed my thinking down” and that unaided tests were “hard to think,” and Miami art-history teacher Kaitlyn Ruano, whose students failed an AP exam after district access to Gemini, recounting a student who said, “I’ve been using AI all year.” The paper reports Google kept pushing the software despite its own researchers warning that children are particularly at risk of mental and emotional harm, and quotes former Google research scientist Mikel Rodriguez, who left last year: “We’re doing this experiment on an entire generation on cognitive development, with no control group.” It argues Google’s Chromebook distribution and network of more than 6,000 certified teacher influencers give it a classroom foothold other chatbots lack.
Our Take (balanced)
The left’s strongest argument is that the skill drop is real but mediated by use pattern: the high-school math retest and UC Irvine offline results show gains reversing without AI, and the augmentation-versus-automation experiments are the best evidence that tutor-like, friction-heavy use can protect independent performance. The right’s strongest argument is that the setting is not a voluntary adult workflow—preloaded school tools, Google’s education-channel reach, students who could not sit unaided exams, and the OECD’s 25-year low plus the 20-point science gap for daily users—pointing to default classroom dependency that individual intentionality may not reach. The central unresolved tension is whether eroded independent scores are mainly a habit problem that better prompting can fix, or the foreseeable result of installing AI as school infrastructure for children who have not yet acquired the skills being offloaded.
2 sources
More stories
The week's bottom lines, in your inbox
One email a week: the five stories that mattered and what they actually mean. Free.