# AI foundations practice workbook

Universal AI Institute — Research. Education. Equity.

Use fictional or shareable public data. Keep the prompts, outputs, and your own review notes.

## A human guide to artificial intelligence

choose one useful task
1. List three small tasks you do regularly.
2. Choose one with non-sensitive inputs and an output you can easily check.
3. State whether you want transformation, exploration, investigation, or a combination.
4. Try one prompt and save the response.
5. Mark every statement that was not present in your input. Decide which additions are helpful suggestions and which are unsupported facts.

**Your finished artifact:** The original input, the AI draft, and a corrected version with a short note explaining your changes.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next:

## What a language model knows—and doesn’t

test two boundaries
1. Run the prompt above.
2. Confirm that the answer rejects an overnight loan and marks the ID requirement as unspecified.
3. Ask the same question in a new conversation without supplying the policy.
4. Compare the responses. Do not treat a guess in the second conversation as knowledge of your policy.
5. Record what information you needed to provide for a reliable answer.

**Your finished artifact:** A two-column comparison labeled “Source supplied” and “Source missing,” with one sentence about the difference.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next:

## Better questions. Better answers.

compare two prompts
1. Write a vague prompt for a low-risk task you actually do.
2. Define three success criteria before generating anything. For an invitation, use factual accuracy, audience fit, and clarity of next steps.
3. Run the vague prompt and save its output.
4. Run a four-part brief for the same task in a fresh conversation.
5. Score both outputs against the same criteria, using 0 for missing, 1 for partly met, and 2 for met.
6. Revise one instruction that addresses the weakest result. Run it again.

**Your finished artifact:** Three prompt versions, their outputs, and a six-point comparison. A better score is useful evidence for this task, not proof that one prompt wins for every task.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next:

## Verify before you trust

audit a short response
1. Use the fictional report or a short public source you can inspect.
2. Ask an assistant for a summary.
3. Choose three factual claims from the output.
4. Verify each against the source yourself. Recalculate any arithmetic.
5. Correct the output and explain the most important correction.

**Your finished artifact:** A three-claim ledger and a corrected summary. Do not mark a claim supported just because another AI assistant agrees with it.


### Your work

Input/source:

Prompt:

Output:

What I checked:

What I changed and why:

What I will try next: