AI Constitution

Resources · Learning tool

Education Concept Tracker

A prerequisite knowledge graph of 1590 learning concepts across 8subjects. Track what you've mastered and see exactly what you're ready to learn next — the topics whose prerequisites you've already met. Built on the Marble Open Skill Taxonomy.

Sign in to track your progress and get personalized recommendations. You can browse freely below.

Knowledge-graph map

Learning to Learn
Mastered Ready to learn LockedHover a node for its description · click to open
Asking for HelpAsking for Help — frontier Ask for help when you've had a go yourself and are still stuck — knowing when to ask is a skill in itselfChecking Your Own WorkChecking Your Own Work — frontier After finishing a task, look back at what you did and ask yourself: does this seem right?Persisting When It's Ha…Persisting When It's Hard — frontier Keep trying when something feels hard — making mistakes and trying again is how learning happensFeeling of not understa…Feeling of not understanding — locked Notice the feeling of not understanding — recognise when something is confusing rather than reading or listening past itPlanning a TaskPlanning a Task — locked Make a simple plan before starting a task: what do I need to do, and what should I do first?Thinking Before StartingThinking Before Starting — locked Before starting something new, stop and think: what do I already know about this topic?Spotting PatternsSpotting Patterns — frontier Spot patterns and recurring structures — in numbers, words, nature, sounds, or events — and use them to make sense of new informationDescribing Rules & Patt…Describing Rules & Patterns — locked When you notice a pattern repeating, describe it as a rule that works every time — then test whether the rule holds in new casesLearning from MistakesLearning from Mistakes — locked When you get something wrong, investigate why — what did you misunderstand or overlook? Analysing errors is one of the most powerful ways to learnTransferring SkillsTransferring Skills — locked Recognise when a skill or strategy learned in one subject or situation can be applied in a completely different one
← Back

Bias in AI Systems

Computing · Artificial Intelligence · conceptual

If training data is biased, AI will be biased; examples: facial recognition working better for some skin tones, translation assuming gender; where bias comes from and whether we can fix it

Sign in to track this

What mastery looks like

  • Explain what bias in AI means using a real-world example
  • Describe how biased training data leads to biased AI results
  • Suggest one way to reduce bias in an AI system (use more diverse data, test with different groups)

Check your understanding

Could Bias in AI Systems explain why an AI trained mostly on photos of light-skinned faces might not work as well for people with darker skin?

Prerequisites

Required first

Unlocks (2 directly)