Locked learning resources

Join us and get access to thousands of tutorials and a community of expert Pythonistas.

Unlock This Lesson

Locked learning resources

This lesson is for members only. Join us and get access to thousands of tutorials and a community of expert Pythonistas.

Unlock This Lesson

You're Still the Thinker

Sarah Jane Ross
AI for Beginners: Use Agents With Confidence Sarah Jane Ross 03:12

By now you know what’s happening behind the chat box and how to check what comes out of it. Checking is one decision that stays with you. This lesson covers the others, starting with what good means for your work.

Treat It as a First Draft

An AI assistant is very good at producing a first version: a draft, a summary, a list of options, a plan. What it can’t do is decide what good means for you. It doesn’t know which detail a client will care about, which of two vendors fits your budget, or whether a proposal sounds like your team.

So treat what it gives you as material to work with. Read it, change it, and keep what’s right. Whatever you send, submit, or act on goes out under your name, and you’re the one who answers for it. That’s why the checking from Habits for Checking Results matters: the AI can’t judge the result for you.

Using It to Learn

Locked learning resources

Join us and get access to thousands of tutorials and a community of expert Pythonistas.

Unlock This Lesson

Already a member? Sign-In

Locked learning resources

The full lesson is for members only. Join us and get access to thousands of tutorials and a community of expert Pythonistas.

Unlock This Lesson

Already a member? Sign-In

00:00 You’re Still the Thinker.

00:03 By now you know what’s happening behind the chat box, and how to check what comes out of it. Checking is one decision that stays with you. This lesson covers the others, starting with what good means for your work.

00:15 An AI assistant is very good at producing a first version. A draft. A summary. A list of options. A plan. What it can’t do is decide what good means for you.

00:27 It doesn’t know which detail a client will care about. Or which of two vendors fits your budget. Or whether a proposal sounds like your team.

00:36 So treat what it gives you as material to work with.

00:39 Read it, change it, and keep what’s right. Whatever you send, submit, or act on goes out under your name. You’re the one who answers for it. That’s why checking matters.

00:51 The AI can’t judge the result for you.

00:54 The same assistant that can do a task for you can also teach you to do it. The difference is in what you ask for. Ask it to explain how it got a result, step by step.

01:04 Ask it to quiz you, to check that you understood. Show it your own attempt, and ask what you got wrong, and why. Or ask for one hint at a time, instead of the answer.

01:15 If you hand every step of a job to an AI, you don’t learn how the job works. That’s fine for things you never need to understand. But for anything you care about getting right, there’s a cost.

01:26 When you can’t do the work yourself, you can’t tell whether the AI did it well. People who build apps by describing them, without ever looking at what was built, often reach this point.

01:36 The app mostly works, and they can’t tell why it breaks. So for the work that matters, stay involved. Make the first attempt yourself now and then. Or ask the AI to explain its choices before you accept them.

01:49 Some jobs are better done without an AI, or with one only in a supporting role. When a decision affects people, money, contracts, or compliance, you make the call.

02:00 For facts you must be sure of but can’t check, go to the source. That might be the official website, the contract itself, or your legal team. Passwords and ID numbers stay out of the chat.

02:11 And some workplaces, clients, and certification exams don’t allow AI for certain work. Their rules come first.

02:19 Three more things come up once you use these tools for real. First, bias. The text a model trained on was written by people, and that text carries people’s assumptions.

02:29 So if an answer is about people, like job candidates or customers, read it with that in mind. Second, who owns what it writes. That depends on the app’s terms and the laws where you live, and those rules are still changing.

02:42 If it matters, check the app’s terms and your company’s policy first. Third, being open about it. Follow your workplace’s and your clients’ rules. And when it would matter to the reader, say that you used AI, and how.

02:56 So the AI makes a first version, and you decide what good means. Below this video, you can decide how much of each work task you’d give an AI assistant. A few tasks have a second good answer, so read the reason for each one.

Become a Member to join the conversation.