My practice workbook

Seven modules to understand, experiment and exercise judgment.

Fill in this document, then use your browser’s Print command to keep it on paper or as a PDF. Fields are neither stored nor transmitted; reloading clears them. This workbook is not a certification.

Back to the course →

Module 1 of 7

Open the machine

Understand what generative AI does and what it does not guarantee.

Trace the four stages: your message, the pieces (tokens), the prediction, your judgment. At which stage does a person verify the facts?

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Module 2 of 7

Prediction is not knowledge

Distinguish probability, variation and accuracy.

Keep the context unchanged. Compare two temperatures. What changes? What does that not prove?

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Module 3 of 7

Follow the evidence

Distinguish a supported, contradicted or unstated claim.

Fictional source: the workshop is proposed for Wednesday at 10 am; the room is unconfirmed. Classify: Wednesday at 10 am; room confirmed; workshop free. Justify each classification.

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Module 4 of 7

Build an instruction

Specify a task, an audience and evaluation criteria.

Write an instruction for a 15-minute activity. Specify audience, format, requested evidence and limits to check.

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Module 5 of 7

Choose what to share

Identify personal information and reduce what you transmit.

Fictional note: the only night-shift lead at branch X has been absent since Monday. Rewrite it using only what is necessary to arrange cover.

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Module 6 of 7

Look for who is missing

See why an average can hide very different results.

A succeeds at 95%, B at 60%. Calculate the average with 80 A and 20 B, then with 50 A and 50 B. Explain the difference without attributing a performance change to either group.

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Module 7 of 7

Decide, then apply

Combine the habits in a complete scenario.

Prepare an announcement from the final project notes. Separate established facts, information to confirm, data to remove and the person responsible for sending.

My observations and reasoningOne action to reuse in my daily work

Fictional examples only. No personal information.

Checkpoints

Review your reasoning

Human judgment checks sources and intended use.
A higher temperature spreads weight across continuations; it does not check facts.
Wednesday at 10 am: proposed, not confirmed. Room confirmed: contradicted. Free admission: not stated.
An assessable instruction specifies task, audience, format and criteria; it does not guarantee accuracy.
Minimal example: arrange cover for a shift.
88%, then 77.5%. Group results stay fixed; their weights change.
90 minutes and 24 places are established in the notes. Date, facilitator and budget remain unconfirmed. The private list is unnecessary. A person checks before sending.

Continue in an in-person workshop with your group. Explore the workshops →