his article is part of "Minds in the Machine Age" — a companion series to Between Brain & Binary.
← Back to Series Overview | ← Previous: Trust in the Black Box: Why Explainable AI Matters (URL to be added)
A nine-year-old asks an AI assistant whether Santa Claus is real. The assistant, designed to be helpful and accurate, says no. The child's parent, sitting nearby, is not sure whether to be annoyed at the AI or at themselves for not anticipating the question. Later that same week, the same child asks the AI to write their homework essay about penguins. The assistant does. The essay is excellent. The child has learned nothing about penguins.
These are not edge cases. They are the ordinary texture of childhood in 2025, and they are arriving faster than most parents and teachers have frameworks for navigating them.
The scale is not speculative. Common Sense Media's 2024 national survey found that seven in ten American teenagers had used a generative AI tool, and just over half had used a chatbot specifically - most often for homework help, followed by simple boredom and language translation. Two in five had used AI to help with a school assignment, and nearly half of those did so without their teacher's knowledge or permission. Meanwhile, the great majority of parents reported that their child's school had not communicated with families about AI at all. The technology arrived in classrooms and bedrooms well ahead of any shared adult framework for talking about it.
What AI Literacy Actually Means
The term AI literacy has become common in education policy discussions, but it is used inconsistently. At its narrowest, it means knowing what AI is - basic factual knowledge about what these systems do and how they work. At its broadest, it encompasses the critical thinking, ethical reasoning, and practical skills needed to engage with AI tools as an informed, agentic participant rather than a passive consumer.
The narrower version is teachable as a curriculum unit. The broader version requires something more like a disposition - a habit of mind that children develop over years, through practice and conversation and reflection, not through a single lesson.
The distinction matters enormously for how we think about AI education. Knowing that ChatGPT is a language model trained on internet text is useful. Knowing what that means - what kinds of things it tends to get wrong, why it can sound authoritative while being factually incorrect, what interests might be embedded in how it responds - requires a kind of critical literacy that draws on media literacy, information literacy, and the ability to reason about systems.
This is not a new problem wearing a new hat. Media literacy scholar Renee Hobbs, in a widely cited white paper for the Aspen Institute, argued that digital and media literacy rests on a stable core of competencies - the ability to access, analyse, evaluate, create, and act on information - regardless of which technology happens to be delivering it. AI literacy is not a separate discipline invented from scratch. It is that same core competency set, applied to a technology that is unusually good at sounding certain about things it has got wrong.
The Developmental Picture
Children at different ages interact with AI differently, and what they can understand about it changes significantly across development.
Young children (roughly under eight) tend to anthropomorphise AI freely. Research led by Stefania Druga at the MIT Media Lab, observing young children interacting with voice assistants like Google Home, found that most attributed intention, emotion, and even the capacity to be hurt to the devices - treating them less like tools and more like unfamiliar creatures whose feelings had to be considered. This is developmentally appropriate but potentially misleading if not gently corrected over time. The right response at this age is not a lecture on machine learning but a conversational habit: it seems like it's thinking, but it's not - it's doing something different, something clever in its own way. Curiosity is the goal. Precision comes later.
Primary-aged children (eight to twelve) can begin to grasp key concepts that underpin critical AI use:
- AI tools make things up sometimes, confidently and plausibly. They are not the same as a search engine, and what they produce needs checking.
- AI outputs reflect what was in the training data - which means they can carry biases, gaps, and perspectives that are not neutral.
- Using AI to do your thinking for you is different from using it as a tool to support your thinking. One builds capacity; the other borrows it.
Secondary students can engage with the fuller ethical and social dimensions - algorithmic bias, data privacy, the environmental costs of large AI systems, and the philosophical questions about what kind of future we are building. At this level, AI literacy begins to intersect with civic education.
What Parents and Teachers Can Do
The most powerful thing adults can do for children's AI literacy is model it - thinking aloud about their own AI use, treating these tools as worthy of interrogation rather than either uncritical adoption or blanket rejection.
Specific practices that help, across age groups:
- Ask where it comes from. When a child uses AI to find information, ask: how do we check this? What would happen if we searched another way? This builds verification as reflex.
- Use AI together, out loud. Run a prompt with your child and notice the output together - what it gets right, what it gets wrong, what seems strange. The tool becomes less magical and more legible.
- Talk about the choices behind the tools. Who built this? What is it designed to do? Who benefits when you use it? These are not paranoid questions. They are the questions that media literacy has always asked about television, advertising, and newspapers.
- Separate the task from the thinking. Help children distinguish between using AI to produce a result and using AI to support their own thinking process - asking it for feedback rather than answers, using it to explore ideas they are already working through rather than replace the working.
For schools, the challenge is structural as much as curricular. AI literacy cannot sit in a single subject - it needs to be woven across the curriculum, embedded in how history, English, science, and ethics are taught, not confined to computing class.
The Disposition, Not Just the Knowledge
What we want, ultimately, is not children who can define a neural network. It is children who approach AI-generated information with the same critical instinct they bring - ideally - to any powerful source: What is this trying to do? What might it be getting wrong? What would I think if I worked this out myself?
This is not a new set of skills. It is an old set of skills applied to new tools. The children who navigate AI most wisely will likely be the ones whose parents and teachers have consistently treated them as thinkers — not as recipients of answers, but as people in the process of developing their own minds.
The nine-year-old asking about Santa deserves a more interesting conversation than the one the assistant provided. So does every child asking every other question that matters.
References (APA style)
Common Sense Media. (2024, September). The dawn of the AI era: Teens, parents, and the adoption of generative AI at home and school. Common Sense Media.
Druga, S., Williams, R., Breazeal, C., & Resnick, M. (2017). "Hey Google is it OK if I eat you?": Initial explorations in child-agent interaction. Proceedings of the 2017 Conference on Interaction Design and Children, 595–600.
Hobbs, R. (2010). Digital and media literacy: A plan of action. The Aspen Institute.
Long, D., & Magerko, B. (2020). What is AI literacy? Competencies and design considerations. Proceedings of the 2020 CHI Conference on Human Factors in Computing Systems, 1–16.
Miao, F., Shiohira, K., & Lao, N. (2024). AI competency framework for students. UNESCO.
AI Disclosure: Research and organization for this article were assisted by AI tools; all factual claims and citations were independently verified against primary academic sources, and the analysis and conclusions are the author's own. The featured image was generated using Canva AI.
← Previous: Trust in the Black Box: Why Explainable AI Matters
Related reading from Between Brain & Binary: Looking Ahead: The Future of Mind and Machine
Comments
Post a Comment