Should AI do this at all?
Practise a thoughtful no as well as a useful, limited yes. This is an original fictional exercise. No tool account, upload or automatic action is needed.
- Read both fictional tasks and name who could be affected.
- Ask whether the result can be checked against reliable inputs.
- Distinguish preparing a limited draft from making a consequential decision.
- Keep the appropriate person and process responsible. Choosing not to use AI is a valid outcome.
Try it here.
Fictional options: A) draft a book-swap invitation from confirmed public details, with a person checking before publishing; B) automatically decide who gets financial assistance, using personal records and no human review.
Both are writing tasks, so give both to an AI tool and accept the result.
Open the explanation without answering
The tasks have different consequences. A person can compare an invitation with confirmed details and decide what to publish. An assistance decision affects people, uses private information and needs a legitimate responsible process, applicable safeguards and appropriate expertise. This exercise does not establish what law applies to any real organisation or certify a workflow.
A little reflection.
Not submitted or saved by the site. Do not forward private reflections. Downloading is optional.
Optional prompt · public or fictional inputs only
Discuss only fictional task choices. Distinguish a checkable public-information draft from a consequential decision. Do not recommend automating payments, eligibility, hiring, health or legal decisions from this exercise. Ask what human process and safeguards would be required.
Pasting a prompt into another service uses that service’s terms. Check settings, age rules and permissions first.
Sources & scope
Edition 2026-10-09.3. AI-assisted editorial revision; independent specialist and reader review pending. Educational material, not legal, medical or financial advice. Examples are hypothetical unless explicitly documented.
How we publish and correct →