Understand what AI does before asking it to do important work.
What you’ll buildCreate and check a useful AI-assisted draft.
AI is a broad name for computer systems that perform tasks we connect with intelligence. Generative AI creates material such as text, images, or sound. It learns patterns from data. A useful answer can still contain a mistake. A confident tone is not proof that the answer is correct.
An app is the tool you use. A model is part of the system that produces the result. A prompt is your request. Begin with a small task: ask for a draft of a short meeting agenda. Give the topic, the people involved, and the result you want. Read the draft and change anything that does not fit.
Learning AI is partly learning how to ask and partly learning how to check. Keep your own goal in view. Do not put private records into a tool just because it can accept them. Check the source when an answer makes an important claim. Start with one tool and one useful task before comparing a long list of products.
An AI tool drafts a meeting agenda and invents a deadline that nobody agreed to.
Pause and choose your first action. Explain why it would help before opening the example response.
Treat the draft as material to review. Check every date against the actual request, remove unsupported details, and record what a person must approve.
Start with plain words, a useful question, and one worked example.
Hear a perspective, notice a choice, and pause to think.
Try a small task. Use feedback to make the next attempt better.
Make something you can use, test it, and reflect on the result.
These six modules outline the full course we are building. Planned topics are not yet full lessons. Start with the published foundation lesson below; available deeper lessons are linked in their relevant modules. No other course is required.
This introductory lesson is ready to use. The six-module curriculum below shows the larger course being developed. Take any course independently.
An app is the interface you use. A model is a system inside or behind it that produces a result from patterns learned during training. Generative AI can produce text, images, audio, or code. It can make useful drafts and convincing mistakes. Different tools have different inputs, limits, and data practices. Choose a tool for a task you understand before comparing many products. A confident answer does not prove that the system checked a fact.
A useful prompt states the goal, relevant context, constraints, and desired format. Instead of “write an email,” explain the audience, purpose, known facts, and tone. Ask the tool to identify missing information rather than invent it. You can request a short draft or a table, then revise the result. Avoid adding private information just to make the prompt more detailed. Use a fictional or approved example when practicing with sensitive work.
Review an AI answer by checking facts, completeness, usefulness, and risk. Compare dates and figures with original records. Open cited sources and confirm that they support the claim. For code, test the behavior; for a plan, test whether it fits the real constraints. Some tasks need a knowledgeable person’s review before the result is used. The important skill is knowing what you can verify and where you need help.
A shop needs an email announcing a changed opening time. The owner supplies the verified date, new hours, audience, and a request for three short sentences. The first draft adds a discount that does not exist. The owner removes the discount and asks the tool to use only supplied facts. They compare the final version with the calendar before sending. The tool helped write the message; the owner supplied the authority and the factual check.
Use fictional details to draft a meeting agenda. Include a goal, audience, known facts, length, and format. Mark each statement as supplied fact, reasonable suggestion, or unsupported addition. Revise the draft and record one limitation.
Ask someone to try your result when practical. Use their experience to revise one choice. The three questions check understanding; the project checks what you can actually do.
Notes stay in this browser when storage is available. Include them in your downloaded learning plan below.
Original InAct explanation and illustrative examples. These primary sources support further study; platform terms should be checked again before use. References use an APA-style organization, date, and title. Sources without a verified publication date are marked n.d.
The written lesson stands on its own. These optional videos offer another example or perspective. Read the focus, watch if you choose, and try the idea yourself. If a video is unavailable, continue with the explanation and practice on this page.
Andrew Ng · Stanford Online
Start at 00:00. Use the talk as a foundation. Write down one task, one possible use of AI, and one result you would need to check. Product examples are from 2023.
Watch on YouTube at 00:00 ↗Stanford Online. (n.d.). Opportunities in AI, 2023 [Video]. YouTube. Original publisher link above. Video and availability remain controlled by the publisher.
Ethan Mollick · The Wharton School
Start at 00:00. Separate demonstrations from claims about every workplace. Sketch a small task where you could compare your work with an AI-assisted draft. This talk is from 2023.
Watch on YouTube at 00:00 ↗The Wharton School. (n.d.). Generative AI & the Disruption of Work [Video]. YouTube. Original publisher link above. Video and availability remain controlled by the publisher.
Videos load only when you choose to watch. The original publisher controls playback; use the direct link if a player is unavailable. Captions and playback speed are available when the publisher provides them.
Your plan stays in this browser on this device. No account or server storage. Download it to keep your own copy.