Using AI Without Being Used By AI
Three Critical-Thinking Skills for Teens.

AI is here and it’s changing the world. Used correctly, AI can be an extraordinary extension of our knowledge, ambition, and imagination. In the history of the world, it has never been easier to go from idea to reality. However, as with any powerful tool, using it correctly comes with real responsibilities. If we aren't careful, instead of us changing the world with the help of AI, it will be AI changing us by distorting our beliefs, impairing our judgement, and even giving us a false picture of reality. The worst part is that we may not even notice these changes because it feels like we’ve been in charge the entire time.
Research shows that people interacting with biased AI systems became more biased themselves (Glickman & Sharot, 2025). In one study, the authors specifically note that participants were “often unaware of the extent of the AI’s influence,” showing how AI can alter human judgment while users do not fully perceive the source of the change (Glickman & Sharot, 2025). Other studies show that when people trust AI too much, they tend to do less critical thinking and they also retain less information. For example, one study showed that when students use AI for learning tasks, the work can feel easier while producing shallower understanding (Lee et al., 2025; Stadler et al., 2024). AI can also be surprisingly persuasive: personalized AI arguments can change people’s views more effectively than human debaters, and even “neutral” AI summaries can silently shift political or social opinions depending on how information is framed (Salvi et al., 2025; Shu et al., 2026).
Learning is an effortful process, not a passive one. It requires repeated focus and directed mental to grow the neural spines and dendrites required for memory consolidation. Without neuroplasticity, changes in thinking, learning, and behavior are limited. If AI is used continuously instead of real mental effort, neurons will fire but they will not wire.
A recent MIT study using EEG brain scans found that students who used AI tools like ChatGPT to write essays had a 47% reduction in active neural connections compared to those writing manually or using traditional search tools. Neural engagement dropped from 79% to 42%, showing a marked decline in cognitive effort. Students relying on AI also showed weaker memory retention, less neural connectivity, and a reduced sense of ownership over their work.
The warning is simple: AI can give people the feeling of knowledge, effort, authorship, and independent judgment without always building the real thing. In an AI world, critical thinking is more important than ever. The solution is not to avoid AI, but to understand how to harness it in ways that protect and empower your own unique thinking.
Use AI like a power tool for your brain: it can help you build more, faster, but you still need to steer it. Each skill below has a teen version and an adult version of the same core habit.
1. Trust But Verify
Dream very big and test very hard. AI lets you dream bigger than ever before. You can build an app, launch a business, write a research paper, compose a film score -- things that used to require years of training or a team of specialists. But the bigger the dream, the better your tests need to be. This starts simple. One of the most common failure modes of AI is their tendency to hallucinate facts. Validating key sources provided by AI and having it check its own work are crucial when building with AI. Your own human judgement is the most important resource at the end of the day.
2. AI is Alien / Map the Jaggedness
AI sounds human-- friendly, confident, and highly intelligent-- while still being a machine producing words from patterns. When talking to other people, we have useful subconscious instincts from thousands of years of evolution and our personal human experience that help us to understand things like whether people are telling the truth what their motivations are.
AI hijacks these instincts by seeming human through language and reasoning. Subconsciously, it is very hard not to treat and trust AI like we would a human being. It is crucial to always remember AI is a completely alien kind of “intelligence.”
While AI is useful and powerful, AI has deep unintuitive weaknesses because of its non-human design that result in unpredictably “jagged” strengths and weaknesses. In one classic example, the same AI outperforming surgeons on medical exams could not count the number of ‘r’s in “strawberry,” since AI uses tokens rather than letters to break up words.
Understanding how AI works and remembering the fundamental differences despite it sounding like a friendly assistant keeps our critical thinking sharp.
3. Dreaming in Detail
The difference between the average AI user and an AI maverick is all in the prompt. Specific, detailed prompting and questioning will increase AI response quality more than anything else. Vague prompts will produce vague results. This is how our critical reasoning is undermined- when we don’t have a clear sense of what we are asking, AI fills in the gaps with its best guess, we will mistake AI assumptions for our own beliefs. Specific, detailed prompts produce personalized results instead of “AI slop.”
Expert AI User Tip: Use audio/microphone prompting in order to speak more easily and naturally to the AI
A Simple Standard for AI Use
-
Before AI: I know what I think, what I want, or what I am trying to find out.
-
During AI: I ask questions, challenge assumptions, and give precise direction.
-
After AI: I can explain, verify, edit, and take responsibility for the result.
Use AI. Don’t let AI use you.
References
Draxler, F., Werner, A., Lehmann, F., Hoppe, M., Schmidt, A., Buschek, D., & Welsch, R. (2024). The AI ghostwriter effect: When users do not perceive ownership of AI-generated text but self-declare as authors. ACM Transactions on Computer-Human Interaction, 31(2), 1–40. https://doi.org/10.1145/3637875
Franklin, B. (2008). The autobiography of Benjamin Franklin. Yale University Press. (Original work published 1791)
Glickman, M., & Sharot, T. (2025). How human-AI feedback loops alter human perceptual, emotional and social judgements. Nature Human Behaviour, 9, 345–359. https://doi.org/10.1038/s41562-024-02077-2
Lee, H.-P., Sarkar, A., Tankelevitch, L., Drosos, I., Rintel, S., Banks, R., & Wilson, N. (2025). The impact of generative AI on critical thinking: Self-reported reductions in cognitive effort and confidence effects from a survey of knowledge workers. Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3706598.3713778
Salvi, F., Ribeiro, M. H., Gallotti, R., & West, R. (2025). On the conversational persuasiveness of GPT-4. Nature Human Behaviour, 9, 1645–1653. https://doi.org/10.1038/s41562-025-02194-6
Shu, M., Karell, D., Okura, K., & Davidson, T. R. (2026). How latent and prompting biases in AI-generated historical narratives influence opinions. PNAS Nexus, 5(3), pgag022. https://doi.org/10.1093/pnasnexus/pgag022
Stadler, M., Bannert, M., & Sailer, M. (2024). Cognitive ease at a cost: LLMs reduce mental effort but compromise depth in student scientific inquiry. Computers in Human Behavior, 160, 108386. https://doi.org/10.1016/j.chb.2024.108386