The Life Starter
Sign up
All courses/Digital

Course 92 · Digital course 17

Understand AI and Machine Learning

Train a simple model, measure its mistakes, and explain where the results do and do not support its use.

Before you start

Prepare paper tables and a calculator or hand arithmetic.

See everything you’ll need
What you’ll leave with

A simple model you can evaluate, with a clear explanation of its mistakes and when human judgment is needed.

Start course Start without an account. Sign in to save lesson progress across devices.

Your project

Start with the practice files. Build your project one lesson at a time.

Practice files are free without an account. Sign in to save your project record in My Proof: a description, checklist, and optional work link. Keep the original files in your own app or notebook; saving a project record does not upload them.

  1. Get your materialsDownload the practice worksheetUse the downloadable worksheet or toolkit alongside the lesson instructions.
  2. Make your projectView the project checklist Work in your own notebook, document, or software. Use the supplied practice examples when you prefer.
  3. Keep a private recordSave a project record Sign in to save a description and checklist, then return to update them.
Course handbook

Know the route before you begin.

Here’s what you’ll need and what you’ll make. You can return to this guide during any lesson.

Who this is for

Learners age 14 and older, including adults exploring technical work or strengthening school and community projects. No employment or university enrollment is needed.

Prepare before lesson one
  • Prepare paper tables and a calculator or hand arithmetic.
  • Use the original fictional box data only; no personal records or AI account is needed.
  • Keep training decisions separate from the supplied evaluation labels and label every result by dataset role.
Your final project

Finish with proof you can use.

Evaluate the fictional box classifier and document its evidence and use boundaries in a model card.

Save a project record in My Proof
See project checklist and review criteria

Keep your working files in your own notebook or software. Save a private description, checklist, and work link in My Proof.

What to include

  • Define the prediction task, baseline and human decision boundary.
  • Document feature/label meaning and independent training/evaluation records.
  • Compare two thresholds and freeze the chosen model.
  • Calculate confusion counts, accuracy, precision and recall against a baseline.
  • Report the separate condition audit and a targeted next evaluation step.

Review and revise

Use these criteria to check your work and choose what to improve.

  • Concepts: features, labels, training and decisions remain distinct.
  • Evaluation: arithmetic and denominators match the original records.
  • Judgment: small samples and uneven performance are not overstated.
  • Accountability: use boundaries and further evidence needs are concrete.
Three ideas to help you along the way
Three useful ideas

The rules behind the steps.

01

Separate prediction from decision

A model’s output does not establish a fact or authorize an action.

02

Keep evaluation independent

Do not use final test outcomes to tune a model while claiming untouched evaluation.

03

Expose the errors

Report relevant error types and conditions, not just one flattering score.

What do these words mean?
Key words

Key terms worth knowing.

Feature
An input available to the model at prediction time.
Label
The reference outcome used to train or evaluate a model.
Training
Selecting or fitting model behavior using examples.
Data leakage
Using information that improperly reveals the answer or compromises independent evaluation.
Precision
The fraction of positive predictions that are actually positive.
Recall
The fraction of actual positive cases the model detects.
Your route

6 lessons.
One useful project.

Work in order the first time. Each lesson makes something the next lesson can use.

  1. 1
    Up next

    Identify what a model is trying to predict

    Distinguish a fixed rule, a learned model and a human decision.

  2. 2
    27 min

    Build a small dataset without leaking the answer

    Define features and labels and separate records by their intended use.

  3. 3
    27 min

    Train and freeze a simple model

    Compare candidate thresholds on training data and lock one before evaluation.

  4. 4
    37 min

    Evaluate errors with the right denominators

    Build a confusion table and calculate accuracy, precision and recall.

  5. 5
    29 min

    Inspect uneven performance and changed conditions

    Compare subgroup errors and describe uncertainty without overclaiming fairness.

  6. 6
    37 min

    Write a model card and human decision plan

    Communicate the model’s task, evidence, limits and responsible use.

Use the tools

Tools to help you practice.

See all tools
Text worksheet

Your course project worksheet

A downloadable text worksheet with each lesson’s working output, space for your answers, and the shared final-project checklist. Save your own copy.

Source transparency

Sources and further study.

Explore the references behind the lessons and take your learning further. Source notes explain access and reuse.

MIT OpenCourseWare — Introduction to Machine Learning
Source & access notes · checked 2026-09-10

Read MIT’s public overview of learning problems, representation, overfitting and generalization. Optional deeper study is more mathematically demanding than this original beginner course.

Access: MIT states the Open Learning Library materials are free to use without enrolling; an optional account tracks progress. This Life Starter course provides no MIT enrollment, credit, certificate or affiliation.

Reuse: Linked for reference. Examples, activities and assessments here are original; free access alone does not grant permission to reuse source text or branding.

Google Machine Learning Crash Course — Dividing Datasets
Source & access notes · checked 2026-09-10

Read train/validation/test separation, representative evaluation and duplicate-example risks. Original box fixtures are used rather than source exercises.

Access: Free public reading; no account or certificate is required for this Life Starter course.

Reuse: Linked for reference. Examples, activities and assessments here are original; free access alone does not grant permission to reuse source text or branding.

Google Machine Learning Crash Course — Accuracy, Recall and Precision
Source & access notes · checked 2026-09-10

Read metric definitions, denominators and threshold trade-offs. Course arithmetic is independently checked on its original eight-row holdout fixture.

Access: Free public reading; no account or certificate is required for this Life Starter course.

Reuse: Linked for reference. Examples, activities and assessments here are original; free access alone does not grant permission to reuse source text or branding.

NIST — AI Risk Management Framework
Source & access notes · checked 2026-09-10

Opened the voluntary framework overview and human/organizational risk-management context. No compliance, fairness certification or production approval is implied.

Access: Free public reading; no account or certificate is required for this Life Starter course.

Reuse: Linked for reference. Examples, activities and assessments here are original; free access alone does not grant permission to reuse source text or branding.

Ready when you are

Start with the first useful move.

You’ll always know what to do, what to make, and where to go next.

Start lesson one