Skip to course content
Free no-code course

AI Tools for Everyday Work / Module 8

Module 8

What AI can and cannot know about a table

Unit ID: AIT-M08-U00 Estimated active time: 15-20 minutes

The main idea

The tool can only use the data or description it actually receives.

This matters because everyday AI tools can produce useful work very quickly. Speed is helpful, but speed can also make us skip judgement.

In this unit, keep the Module 1 rule in mind:

AI can assist. You still review and decide.

Why this matters

State whether the tool has the real data, sample rows, or only a description.

When you give clear instructions and review the result, the tool becomes easier to use responsibly. When you skip the habit, the tool may produce something polished but unsuitable.

Safe example

If you describe a sales table without rows, the tool cannot calculate totals.

This is a safe learning example because it does not require private, confidential, personal, customer, student, employee, legal, medical, financial, or sensitive information.

Better working habit

State whether the tool has the real data, sample rows, or only a description.

Write the habit down before you use the tool. A short habit is easier to remember than a long rule.

What to watch for

Do not let the tool pretend to analyse data it has not seen.

If you notice this problem, slow down. You can edit the prompt, ask a follow-up question, review manually, or reject the output.

Mini practice

Write three versions: real data seen, sample seen, description only.

Write your answer in 3-6 sentences or a short table. Keep it safe and nonsensitive.

Check yourself

Before moving on, answer:

  • What did I ask the AI tool to help with?
  • What did I keep responsibility for?
  • What would make this task unsafe or unsuitable?
  • What would I check before using the output?