AI Made It vs. I Made It: Why the Difference Matters for Kids

A 9-year-old holds up a dragon picture on a tablet and says, “Look what I made.” Then, a beat later: “Well, the AI made it. I told it what to do.” That correction is the whole topic. The gap between AI made it vs. I made it isn’t about honesty or rules. It’s about what a child gets from doing the work, and decades of psychology research are surprisingly specific about what that is. Kids value things more when they built them. They remember what they produced better than what they read. Pride in their own work predicts later achievement. And the strongest source of the belief “I can do this” is having done it. AI can produce a better-looking dragon in four seconds. What it can’t produce is the child’s labor, and nearly every benefit above runs through the labor. This article lays out the four pillars of that research, what each predicts about AI, where the evidence is thin, and how to keep the “I made it” in your house.

Key Takeaways

• Ownership: Preschoolers transfer ownership of an object to whoever invested creative labor in it, more strongly than adults do (Kanngiesser, Gjersoe and Hood, 2010).
• Memory: People remember what they generate better than what they read, the “generation effect” (Slamecka and Graf, 1978). In a 2025 MIT preprint, adults who wrote essays with an AI struggled to quote their own work.
• Pride: In 3,425 German students followed from grades 5 to 9, pride and enjoyment predicted later math achievement, and achievement predicted pride (Pekrun et al., 2017).
• Self-efficacy: Mastery experience is “typically the most influential source of self-efficacy” (Usher and Pajares, 2008). A success the AI produced isn’t a mastery experience.
• Adults using AI text often don’t feel they own it, yet still don’t credit the AI (Draxler et al.), and 38% of teen AI users told Common Sense AI leads them to come up with fewer of their own ideas.

AI Made It vs. I Made It: What the Difference Actually Is

Authorship, in the sense that matters for kids, means doing the generative part of the work yourself: deciding, drafting, building, fixing. “AI made it” means a model did the generating while the child described what they wanted.

That isn’t automatically bad. A child who uses AI to look up how dragons are drawn, then draws one, made it. A child who types “draw a dragon” and gets one didn’t. Most real cases fall in between, which is why the useful question is not “did you use AI?” but “which part did you do?”

The research below doesn’t study AI and kids directly in most cases. It studies what happens when people do or don’t do the work themselves. That’s a strength, because the findings are old and well replicated. It’s also a limit: applying them to AI is an inference, and we’ll flag where direct evidence exists and where it doesn’t.

Pillar 1: Ownership Comes From Labor

In a 2010 study in Psychological Science, Patricia Kanngiesser and colleagues at the University of Bristol had an experimenter and a participant borrow clay objects from each other and mold them into new ones. Participants “were more likely to transfer ownership to the second individual after he or she invested creative labor in the object” than after simply holding it or making small changes. The effect “was significantly stronger in preschool children than in adults.”

Young children have a clear, intuitive rule: the person who did the creative work owns the thing. That’s the rule your 9-year-old applied to the dragon.

Adults show a version of it with AI. In two studies (30 and 96 participants), Fiona Draxler and colleagues found what they named the “AI Ghostwriter Effect”: “Users do not consider themselves the owners and authors of AI-generated text but refrain from publicly declaring AI authorship.” The more influence participants had over the text, the more ownership they felt. The study appeared in ACM Transactions on Computer-Human Interaction.

Pillar 2: You Remember What You Produce

In 1978, Norman Slamecka and Peter Graf ran five experiments comparing memory for words people generated themselves with the same words simply read. “In all cases, performance in the generate condition was superior to that in the read condition,” they reported in the Journal of Experimental Psychology: Human Learning and Memory. They concluded “the generation effect is real.” We cover the classroom side in our explainer on the generation effect.

When AI generates, the child is in the “read” condition.

There’s now direct, if early, evidence for this with AI. In “Your Brain on ChatGPT,” an MIT Media Lab preprint by Nataliya Kosmyna and colleagues, 54 adults wrote essays with an AI chatbot, a search engine or no tools. “Self-reported ownership of essays was the lowest in the LLM group and the highest in the Brain-only group. LLM users also struggled to accurately quote their own work.” The study is not peer-reviewed, involves adults, and has drawn a published methodological critique, which we cover in our article on the MIT study. But the memory finding is exactly what the generation effect would predict.

Pillar 3: Pride Feeds Achievement, and Achievement Feeds Pride

Reinhard Pekrun and colleagues followed 3,425 German students from grades 5 to 9 in the PALMA study, published in Child Development in 2017. Positive emotions such as enjoyment and pride “positively predicted subsequent achievement,” and “achievement positively predicted these emotions,” even after controlling for gender, intelligence and family socioeconomic status.

That’s a loop: doing well produces pride, pride fuels effort, effort produces doing well. The research doesn’t test AI, but it’s fair to ask what happens when a kid’s best-looking work isn’t theirs. Pride in something you didn’t make is fragile. Kids often know it, which is why the 9-year-old corrected himself.

The IKEA effect sits nearby. Michael Norton and colleagues found adults valued their own amateur Lego and origami creations about as highly as experts’ work, but only when they finished them. Lauren Marsh and colleagues found the bias emerges around age 5 and may reflect “a link between our creations and our self-concept.” We go deeper in the IKEA effect in kids.

Pillar 4: “I Can” Comes From “I Did”

Albert Bandura’s theory of self-efficacy names four sources of the belief that you can succeed at a task. In a review in the Review of Educational Research, Ellen Usher and Frank Pajares concluded that across school research, “mastery experience is typically the most influential source of self-efficacy,” though its strength varies by gender, ethnicity, ability and subject.

Mastery experience means succeeding at something yourself, ideally something that was hard. An AI-produced success can be satisfying in the moment. It isn’t mastery, because the child didn’t do the thing that would let them believe they could do it again.

Teens seem to sense this. In Common Sense Media’s 2026 survey of more than 1,000 US teens, Teens in the AI Era, 38% of teen AI users said AI leads them to generate fewer of their own ideas, and 39% said they feel they’re missing out on learning when they use AI to complete assignments.

The Four Pillars at a Glance

Pillar	Key study	What it found	What changes when AI does the work	What to protect
Ownership	Kanngiesser et al., 2010	Kids assign ownership to whoever did the creative labor	The child did little labor, so it doesn’t feel theirs	Let the child do the generative step
Memory	Slamecka & Graf, 1978; Kosmyna et al., 2025 (preprint)	Generating beats reading; AI essay writers couldn’t quote their own work	The child is reading, not generating	Draft first, AI second
Pride	Pekrun et al., 2017	Pride and achievement feed each other over years	Pride in borrowed work is fragile	Praise the part they made
Self-efficacy	Usher & Pajares, 2008	Mastery experience is the strongest source of “I can”	A success the AI produced isn’t mastery	Hard, finishable tasks done solo

A fair caveat on the whole table: only the memory row has direct AI evidence, and that evidence comes from a small adult preprint. The rest are well-established findings applied to a new situation. That’s a reasonable inference, not a proven effect.

Where AI Help Doesn’t Cost the “I Made It”

None of this argues for keeping kids away from AI. It argues for keeping the generative step theirs.

AI before or after, not instead of

Using AI to explain a concept before a child writes, or to give feedback after they’ve drafted, leaves the making with them. Using it to produce the draft moves the making to the machine. Draxler’s finding, that more influence means more ownership, points the same way.

Praise the part they made

When your child shows you something made with AI, ask: “Which part was yours?” Then praise that part specifically. You’ll reward the labor, and you’ll get an honest answer more often.

Make room for finished, imperfect work

The IKEA effect needs completion. A lopsided clay dragon that’s finished does more for a kid than a perfect AI dragon. Display the lopsided one.

Ask for the story behind it

“How did you make this? What was hard?” Kids who did the work can answer at length. Kids who didn’t usually can’t, and they notice the difference too.

What not to do

Don’t shame a child for using AI or treat every AI-assisted project as cheating. That pushes use into hiding and teaches nothing about where the line is. The goal is for your child to feel the difference between “AI made it” and “I made it,” not to fear the tool.

What to Watch For Over the Next 3 Months

• Week 4: When your child shows you a project, can they tell you which parts they did and how? Have you asked “which part was yours?” at least a few times?
• Month 2 red flags: Your child can’t explain their own homework, seems indifferent to work they “made,” or says things like “there’s no point, AI does it better.”
• Month 3 self-check: Is there something on your fridge, shelf or wall that your child made entirely themselves this term? If not, make space for a project where AI isn’t part of the making.

Frequently Asked Questions

Why does it matter if AI made my child’s project?

Because most of the benefits of making things come from the labor, not the result. Research links doing the work yourself to feeling ownership, remembering better, building pride that feeds later achievement, and believing you can succeed. When AI does the generating, the result can look better while those benefits shrink.

Is any AI use okay for creative projects?

Yes. AI that explains, inspires or gives feedback, while the child does the drafting, drawing or building, leaves the authorship with the child. A useful test is asking which part your child did and whether they could do it again without the AI.

Is there research on kids, AI and ownership specifically?

Not much yet. The strongest direct evidence is from adults: AI users felt less ownership of AI-generated text and, in one MIT preprint, struggled to recall their own AI-assisted essays. The research on children shows they assign ownership to whoever did the creative labor, which suggests the same pattern.

How do I talk to my child about using AI on art or writing?

Ask what they made and how, not whether they used AI. Praise the parts they created. Explain that the tool can help, but the part that grows their skills is the part they do themselves. Most kids already sense the difference.

Will my child fall behind if they don’t use AI to make things?

Unlikely. Using AI well depends on knowing what good work looks like and being able to judge and fix it, skills built by making things yourself. A 2025 Microsoft Research survey of knowledge workers found higher self-confidence went with more critical thinking when using AI.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Kanngiesser, P., Gjersoe, N., & Hood, B. M. (2010). “The effect of creative labor on property-ownership transfer by preschool children and adults.” Psychological Science, 21(9), 1236–1241. https://doi.org/10.1177/0956797610380701
2. Slamecka, N. J., & Graf, P. (1978). “The generation effect: Delineation of a phenomenon.” Journal of Experimental Psychology: Human Learning and Memory, 4(6), 592–604. https://doi.org/10.1037/0278-7393.4.6.592
3. Kosmyna, N., Hauptmann, E., Yuan, Y. T., et al. (2025). “Your Brain on ChatGPT: Accumulation of Cognitive Debt when Using an AI Assistant for Essay Writing Task.” arXiv (preprint, not peer-reviewed). https://arxiv.org/abs/2506.08872
4. Pekrun, R., Lichtenfeld, S., Marsh, H. W., Murayama, K., & Goetz, T. (2017). “Achievement Emotions and Academic Performance: Longitudinal Models of Reciprocal Effects.” Child Development. https://doi.org/10.1111/cdev.12704
5. Usher, E. L., & Pajares, F. (2008). “Sources of Self-Efficacy in School: Critical Review of the Literature and Future Directions.” Review of Educational Research. https://doi.org/10.3102/0034654308321456
6. Draxler, F., Werner, A., Lehmann, F., Hoppe, M., Schmidt, A., & Buschek, D. (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. https://doi.org/10.1145/3637875
7. Norton, M. I., Mochon, D., & Ariely, D. (2012). “The IKEA effect: When labor leads to love.” Journal of Consumer Psychology. https://doi.org/10.1016/j.jcps.2011.08.002
8. Marsh, L. E., Kanngiesser, P., & Hood, B. (2018). “When and how does labour lead to love? The ontogeny and mechanisms of the IKEA effect.” Cognition, 170. https://doi.org/10.1016/j.cognition.2017.10.012
9. Common Sense Media. (2026). “Teens in the AI Era: Schoolwork and Skills That Matter.” https://www.commonsensemedia.org/research/teens-in-the-ai-era-schoolwork-and-skills-that-matter
10. Lee, H.-P., Sarkar, A., Tankelevitch, L., et al. (2025). “The Impact of Generative AI on Critical Thinking.” CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3706598.3713778