Knowledge Strategy
Knowing how you know what you know — and deciding what's worth keeping in your own head.
What this skill is
Knowledge strategy is metacognition about your own knowledge: where your beliefs come from, how confident you should be in them, and — crucially — what you need to actually know versus what you can safely delegate to a tool.
In a pre-AI world, knowledge strategy mostly meant: study this, remember that, use reference books for the rest. In an AI world, the question becomes more interesting: given that I can retrieve almost anything on demand, what's worth building into genuine understanding? What knowledge has to live in me because I'll need it in contexts where retrieval isn't possible — under pressure, in conversation, in moments of judgment?
A child with good knowledge strategy doesn't just know things. They know which things they know, why they know them, how reliable their sources were, and what they'd need to check before betting on them.
Why it matters in an AI world
AI retrieval is so good that it tempts children (and adults) to stop internalizing knowledge entirely. Why learn it when you can look it up? The answer isn't nostalgic — it's practical. Knowledge that lives inside you enables a kind of thinking that knowledge you retrieve does not. When a doctor knows anatomy deeply, they reason about edge cases in ways that a doctor consulting a reference manual cannot. When a programmer deeply understands algorithms, they debug differently than someone who always turns to search.
There's also the reliability problem. AI is wrong in ways that are hard to detect, and the child who has no internal reference — who has outsourced their knowledge base to AI — has no alarm bell when AI gives them a confident incorrect answer. The child who actually knows things notices the discrepancy. That internal knowledge base is a quality-control mechanism, not just a repository.
Finally, knowledge strategy includes knowing where your knowledge came from and how trustworthy that source is. "I know this because I read it in a peer-reviewed paper" and "I know this because AI told me" are different epistemic states with different levels of warranted confidence. A child who can't distinguish between them is epistemically blind.
What it looks like in your child
- They can distinguish between what they genuinely understand and what they can just retrieve — and they know which is which for a given topic
- When learning something new, they have a view on whether they need to internalize it or just know where to find it
- They can identify the source of a belief and give a rough reliability estimate ("I read this somewhere but I'd need to verify it before I'd stake anything on it")
- When AI gives them something that contradicts what they thought they knew, they don't immediately defer — they investigate the conflict
Challenge: Try this this week
The Source Audit. Pick a topic your child knows reasonably well — a historical event, a scientific concept, something from a class. Ask them to list five things they believe about it and, for each one, name the source (book, teacher, AI, YouTube video, own experience). Then rate each source: how reliable is it, and would you feel comfortable defending this claim to someone who knew the field? This is the full knowledge-strategy workout: not just what you know, but how you know it and how much you should trust it.
What to watch for
- AI as the only source: Everything they "know" traces back to AI — they've stopped reading books, talking to experts, or forming views from experience. AI becomes the single pipe into their knowledge base.
- No distinction between internalized and retrieved: They can't tell you the difference between understanding how photosynthesis works and being able to look up how photosynthesis works. Both feel like "knowing."
- Certainty without provenance: They state things confidently without being able to say where the belief came from. This is often a sign that the belief was absorbed passively — from AI, from social media — without any epistemic processing.
Games that develop this skill
Clue Hallucinated Evidence — In a Clue game, shuffle in a small number of "AI Evidence Cards" that describe clues that aren't actually in the game (invented suspects, rooms, weapons). Players must figure out which evidence is real and which is hallucinated. Forces active evaluation of source reliability mid-game.
Scrabble Noise Rack — Before each turn, a player draws one additional "noise tile" that they must discard before playing — but they don't know in advance which tile is noise. After discarding, they must explain why they chose that tile to discard. Develops the habit of identifying what doesn't belong in your information set.
Where Did I Learn That? — A simple dinner table game. Someone states a fact. Everyone tries to trace it back to its source as specifically as possible. The person who gets the furthest wins. ("I know the Earth orbits the Sun because... my teacher told me in third grade... but she probably got it from... a textbook... which got it from...") This builds awareness of knowledge chains.
See all nine skills
Each skill builds something distinct. Browse the full set to see where your child might have gaps.
All Skill Builders →