Discovery & Questioning
Observe the world around you, find a problem worth investigating, and shape it into a question that can guide real inquiry.
Mission
Being a scientist
is not a job,
but an action.
The goal of PBL (Project-Based Learning) is not to finish a project. It is to experience thinking like a scientist. AI's ultimate value is giving every student the capacity to enter that experience.
S0Discovery Phase
What makes antibiotic resistance spread faster in schools?
AI Coach feedback
“Strong hypothesis. Which variable will you control first?”
Phase gate unlocked
Planning & Research Design
The Four Phases
From first question to final defense
Planning & Research Design
Clarify methods, timeline, and sources. The AI Coach prompts every important decision without writing a word for you.
Execution & Data Collection
Run the investigation, record each step, and collect real data. The AI Coach stays alongside you as scaffolding, never a substitute.
Finalisation & Defense
Write your report, organise process evidence, and prepare for defense. Here the AI Coach takes a step back: the work is yours.
The AI Harness
LearnPBL's AI isn't here to do the project for you. It's here to make sure you do it.
It asks. You answer.
The AI Coach operates in Socratic mode at every phase and decision point. Hardcoded red lines prevent it from writing your report, conclusion, or references. If data fabrication is implied, the coach intervenes.
You can't skip ahead.
Fourteen gate conditions govern phase progression. Planning must be complete before execution unlocks. Content must meet quality thresholds before you advance. The AI cannot bypass this — the system enforces what a good mentor would enforce.
Your work is verifiable.
Every project generates a cryptographically signed Academic Integrity Dossier — combining your declaration, AI-usage provenance audit, and citation verification against Crossref. When universities or judges ask whether a student actually did this, there is a signed answer.
Why This Matters
The most successful practitioners in every field learned by doing. Surgeons practice on cadavers, engineers build bridges, writers revise manuscripts. Online education has defaulted to the least effective method: watching someone else do it.
LearnPBL is built around three decades of learning science — constructivist theory (Piaget, Vygotsky), retrieval practice (Roediger), and deliberate practice (Ericsson) — all pointing to the same conclusion: we learn by doing, reflecting, and doing again.
The AI is the scaffold. The project is the learning. You are the author.