Guide
Getting Useful Help from AI
Turn a vague request into a task an assistant can work on, then check and improve its reply.
Making Sense of AI · Part 5 of 8
Updated · 3 min read
https://greglum.com/learn/getting-help-from-ai
Have a notice to shorten, a list to organize, or a paragraph to draft? Start with a task where you can recognize a useful result. The assistant gives you a draft to work on, and you decide whether it helped.
You do not need special commands. Name the job, give the relevant information, and describe the result you want. Then read the answer before asking for changes.
Consider this fictional notice, written for this lesson:
"Book swap: Saturday, November 14, from 10 a.m. to noon at the community hall. Bring up to 5 books. Entry is free. Children must come with an adult."
"Make this better" leaves a lot open. Better for whom? Shorter, friendlier, or more detailed? A more useful request would be:
"Turn this notice into a short invitation for a neighborhood group. Keep the date, time, location, book limit, cost, and rule for children. Use only the information provided. Do not add arrangements that the notice does not mention."
Now the assistant has a purpose, source material, and boundaries. You can check the invitation against them, even if it misses an instruction.
Read the draft against the original. Did "up to 5" become "at least 5"? Did the rule about children survive? Did refreshments appear somewhere along the way? A friendly invitation can become a different event if you let the details drift.
You can ask for a specific revision: "Keep the adult requirement. Shorten the introduction to 1 sentence." You can also ask it to point out missing information, such as whether people need to register. Take those unanswered questions to the organizer.
OpenAI's ChatGPT fundamentals describes starting with a request and refining it through follow-up questions. It is a vendor's product guide, with workplace examples. The same conversational approach can help with ordinary tasks, without adopting its business scenarios.
Useful starting points include rewriting your own nonprivate paragraph, organizing a fictional packing list, or explaining an unfamiliar term in simpler words. Ask for alternatives when there is no single right answer: "Give me 3 possible titles and explain how their tones differ."
For a simple calculation, a calculator may be easier to check. For an event's opening time, the organizer's current notice is the better authority. For a difficult personal decision, a conversation with someone who knows the situation may provide context an assistant lacks.
An assistant can help you prepare questions for those sources. A polished answer does not remove the need to consult them. OpenAI's truthfulness guidance acknowledges incorrect information and invented references, even in confident responses.
Try improving this fictional request before reading the example below: "Help me pack."
The traveler is taking a 2-night train trip, has 1 small bag, and can wash clothes at the destination. They want a short packing list, not shopping recommendations. What would you tell the assistant?
One possible request: "Suggest a packing list for a 2-night train trip with 1 small bag. I can wash clothes there. Group the list by purpose, avoid shopping recommendations, and ask me about anything that would change your suggestions."
The request sets the task, limits, and format, but leaves weather, planned activities, and access needs unknown. A list with hiking boots and evening wear might fit someone else's weekend perfectly. Ask for the missing details before treating the list as yours.
For a real task, try a different setting: organize your own nonprivate study notes, compare ideas for a volunteer meeting, or rewrite a message in a clearer tone. Decide what a helpful result would look like first. Use that goal to assess the draft and choose which suggestions to keep.
Outside courses that cover this ground in more depth. Compare the time, cost, and limits before you start, or see every next step.
AI Fluency: Framework and foundations
Learn a method for working with an assistant
Anthropic · Claude Academy
- Suits
- People who want a structured way to decide what to hand to AI and how to judge the result.
- You'll Learn
- Choosing tasks for AI, describing them, judging the output, and taking responsibility for how you use it.
- Before You Start
- The course page lists none.
- Time
- 14 lessons, about 4 hours, with a final assessment.
- Cost & Access
- Free. Start without an account; sign in to save progress and earn the badge.
- Keep in Mind
- Written by two university educators with Anthropic and hosted on Anthropic's site. It aims to hold up whatever tools come next.
- Suits
- People who want guided practice using AI at work and a certificate they can show an employer.
- You'll Learn
- Five short courses: AI basics, everyday work tasks, writing prompts, using AI responsibly, and keeping up as tools change.
- Before You Start
- No technical experience.
- Time
- Google says under 5 hours. Coursera's listings for the five courses add up to about 8.
- Cost & Access
- Paid through Coursera: $49 a month in the US and Canada after a 7-day free trial; prices vary by country. Financial aid is available. Needs a Coursera account.
- Keep in Mind
- Practice centers on Google's Gemini. The free part is the trial, not the course.