AI Literacy
- Fan-out
- Split one task into smaller questions you can answer separately.
- See the project notes branch into three questions
- Fan-in
- Bring the separate answers together into one useful result.
- See One short update in the article
- Planner and executor
- Decide what needs to happen before doing it. The planner proposes the work, and the executor carries out only the approved steps.
- See the plan get approved before the draft
- Retrieve and verify
- Find the answer, then check it against the source before you repeat it.
- See the project timeline checked against its sources
- See an answer checked against a guide in AI 101
- Quality gate
- Check the draft against the rules you set before it goes out.
- See the meeting recap checked before it’s sent
- Hallucination
- When AI presents invented or incorrect information as if it were true.
- Office example AI says the kickoff happened on March 17, even though the project email only proposed that date. The status update shows it actually happened on March 24.
- Try this Open the source and check whether the quoted line supports the claim.
- See the fictional example in the article
- Context
- The details AI needs to understand who the work is for, what you want, and what information it should use.
- Office example Giving AI your three notes and telling it the update is for your team.
- See the team update example in AI 101
- Prompt
- What you ask AI to do.
- Office example “Turn these three notes into a short update for my team.”
- See the team update example in AI 101