Kids build their ethical views and trust in artificial intelligence by actually using the tools, not by sitting through formal AI classes or workshops. Practical, hands-on exposure drives how young people evaluate AI credibility, while classroom instruction barely moves the needle.
Direct hands-on experience with AI tools shapes a learner's trust and ethical mindset far more than formal coursework ever will.
If you want to know how your child actually views AI—whether they blindly trust its answers or understand its flaws—don't look at their school syllabus. Look at their screen time.
Signing kids up for theoretical AI workshops or ethics lectures won't automatically make them critical consumers. A child's actual habits using large language models day-to-day dictate how cautious, interested, or reliant they become.
Schools are rushing to build "AI literacy" units, assuming lectures on ethics will prepare students for an automated future. But educators face a growing gap between what students learn in theory and how they behave when prompting a real chatbot. Researchers wanted to isolate what actually drives a student's perspective: formal coursework, self-described familiarity, or raw usage frequency.
Actual daily interaction with AI tools dictates how people view the technology, while classroom instruction leaves almost no measurable footprint.
- Usage frequency drove every perception metric. How often someone used AI tools reliably predicted all five measured outcomes, including their trust in AI accuracy and their desire for further ethics training.
- Formal AI classes showed no effect. Prior attendance in AI workshops or courses had zero statistically significant link to any baseline ethical perspectives.
- Self-assessments were misleading. Asking people how "familiar" they felt with AI was far less predictive of their actual attitude than simply counting their days of use.
- A clear threshold exists. Moving from zero usage to even occasional hands-on interaction triggered a sharp jump in how much students trusted the technology.
Lectures sell the illusion of digital literacy, but tool mechanics teach the real lesson. When kids interact directly with chatbots, they quickly form opinions about what the technology can and cannot handle. Sitting through a presentation about algorithmic bias rarely registers until a student prompts a tool themselves and watches it hallucinate a fact or make a mistake.
This study examined a narrow group of 93 bioscience graduate students and postdoctoral trainees enrolled in a required ethics course. It is also a preprint that has not yet undergone formal peer review, relying on self-reported usage data. While the gap between passive instruction and active practice is striking, we cannot assume an eight-year-old absorbs AI interactions in the exact same way a medical researcher does.
- If you want your teen to understand AI limitations… sit down and test a chatbot together on a homework question to highlight where it makes mistakes, rather than sending them to an online lecture.
- If your child has never touched a generative AI tool… expect their trust in its accuracy to jump significantly after their very first few uses, making early co-use critical.
- If you are trying to gauge your child's AI literacy… ask them how often they open these tools each week rather than what topics their computer class covered.
Skip the passive lectures and focus on guided practice. If you want your child to develop a healthy, skeptical relationship with AI, shape their hands-on habits instead of relying on a school curriculum to do it for you.
Yongkyung Oh, Lynn Talton, Alex Bui (2026). Engagement Intensity as a Learner-Modeling Signal for Adaptive AI Ethics Instruction. arXiv (preprint). — arxiv.org