Prompt

In plain language

A prompt is what we give an AI system when we want it to do something. We might type it, speak it, attach an image or file, or combine several forms of input.

A prompt can include more than a question: it may name the task, provide useful context, set boundaries, show examples, or specify the desired form of the answer.

Why it matters

A prompt shapes the output, but it does not guarantee truth, completeness, or good judgment.

Learning effective prompting practices helps us state goals, provide useful context, set boundaries, and recognize when the result needs verification. The linked Learning Lab modules explore those practices in more depth.

Example

> Turn this fictional announcement into a short checklist. Use only the details provided. Put missing information under “Questions to Ask” instead of guessing.

Common misconception

A prompt is not a magic spell that guarantees a correct result. It steers the system; the resulting output may still need verification, revision, or human judgment.