03 / Independent use / Lesson 6 of 8

Improve a draft without losing facts

Revise an output while preserving its meaning and uncertainty.

20 minutes, including practiceGo to practice Previous lesson

01 / Understand

Decide what must survive

Before rewriting, mark the facts and limits that cannot change. This includes uncertainty: 'may arrive' is not the same as 'will arrive'. A more polished sentence is worse if it quietly changes a promise, deadline or condition.

Ask for a focused revision

Choose a single improvement such as shorter sentences, a calmer tone or a clearer structure. Keep the source beside the revision. Ask the assistant to preserve dates, quantities and qualifying language. Then check those details yourself. Revision is not automatically verification.

Keep the original

Save your original separately before editing a real document. Compare versions and make the final choice yourself. When two drafts make different claims, return to the source; do not choose the version that sounds more professional just because it is smoother.

02 / Worked example

Look closely at the result

A written teaching example, not a live AI response.

Request or source

Fictional source: The parcel may arrive on Tuesday. We will confirm the delivery date on Monday.

Sample output

Your parcel will arrive Tuesday. We guarantee it.

Compare with the tutor’s review

The revision invents a guarantee and removes the confirmation step. A clearer version is: 'Your parcel may arrive on Tuesday. We will confirm the date on Monday.'

03 / Your turn

Put it into practice

Rewrite the source in a friendly tone without changing its meaning. Write one sentence explaining why the sample output is unsafe to send.

Your working notes

Use fictional information. Nothing you write here is sent to a server or an AI. Download your work before leaving; it is not saved automatically.

Review your workI preserved 'may' or equivalent uncertainty.I kept the Monday confirmation step.I made no new guarantee.

04 / Check your understanding

One decision before you go

Take this with you

Improve the wording, then recheck the meaning.

Complete the practice, review each criterion and check your answer to finish your self-review. Every lesson remains open.

Next: Use AI to support your learning

Sources & course notes

Reviewed 5 September 2026. Examples and exercises are original Aulafy teaching material under MIT. These references inform the course’s approach to reliability and learning; they are not endorsements.

NIST: Generative AI risk profileBackground on unreliable outputs and evaluating risks.
UNESCO: Generative AI in educationBackground on human agency, privacy and learning.

Report a correction on GitHub