The Learning Zone

The gap between what you can do alone — and what you can do with help.

Vygotsky's idea: that gap is where learning actually happens.

The Zone is challenging — but achievable with support.

Too easy and nothing sticks; too hard and you quit. That middle band is where you grow.

Vygotsky's “help” was a teacher, a mentor, a peer. Now there's a new helper: AI.

The question is which kind of helper it becomes.

AI can scaffold — get you just past what you could do alone, then fade as you get it. You keep the skill.

AI can collapse the zone — just do it for you. You skip the struggle and never gain the skill.

Same tool, opposite outcome. That's the line we have to find.

Just a theory? MIT tested it.

54 people wrote essays wearing brain sensors, split three ways: ChatGPT · Google · brain-only.

• EEG — sensors on the scalp that read your brain's electrical activity. No needles.
• Brain connectivity — how much parts of your brain talk to each other while you work. More = you're doing the mental work yourself.

The more help you had, the less your brain fired.

• Brain-only — strongest connectivity.
• Google — in the middle.
• ChatGPT — weakest.

Less connectivity = your brain coasting on this task — NOT “brain damage” or “getting dumber.” It's the zone collapsing, measured.

~83%

of the ChatGPT group couldn't quote the essay they'd just written — the least ownership (whether the work actually feels like yours).

They named the cost cognitive debt — easy now, a hidden cost later to memory and independent thinking. No cognitive credit card.

Base first → AI amplifies. Base skipped → AI hollows out.

The group that built the base without AI first, then added it, engaged more — not less. That's scaffolding: support that pays off once you have a base. Exactly the ZPD.

Don't take it as gospel.

• A preprint — not yet peer-reviewed.
• Small sample — 54 people.
• One task — essay writing.
• Other studies show AI can aid short-term; the debate is long-term unaided skill.

But notice: the evidence and the theory point the same way. That skepticism? It's our read-and-judge ethic in action.

Good AI keeps you in the zone — help that fades. Bad AI deletes the zone.

Vygotsky (theory) and MIT (evidence) agree. That's the line our policy has to draw.