Stop Taking AI Courses and Start Using the Tools

You've signed up for a course, haven't you?

Maybe it's on Coursera. Maybe it's a LinkedIn Learning path. Maybe it's one of those £2,000 bootcamps that promises to make you "AI-ready" in eight weeks. Maybe your company enrolled you in something with "fundamentals" in the title.

i'm going to say something that will upset the entire online education industry: most AI courses are a form of productive procrastination. They feel like progress. They are not progress. They're the adult equivalent of highlighting your textbook in three colours and feeling like you've studied.

I know this because i spent four months taking AI courses before i started actually using AI tools. Those four months taught me vocabulary. The first week of actually using the tools taught me everything else.

The course trap

Here's what happens.

You feel anxious about AI and your career. That's legitimate. So you look for something constructive to do. Taking a course feels constructive. It feels responsible. It feels like the kind of thing a sensible professional does. "I'm investing in my development." "I'm staying current." "I'm being proactive."

You sign up. You watch the videos. You do the quizzes. Module 1: Introduction to Artificial Intelligence. Module 2: History of Machine Learning. Module 3: Types of Neural Networks. Module 4: Ethics and Bias in AI.

Eight weeks later, you've got a certificate. You know what a convolutional neural network is. You can explain the difference between narrow and general AI at a dinner party. You feel informed.

And you still can't use AI to do your job better than you could before the course.

Because knowing how AI works and knowing how to work with AI are completely different skills. The course taught you the first one. Your career depends on the second one.

Why doing beats studying

Think about how you learned to use a spreadsheet. Did you take a 40-hour course on the theory of tabular data computation? Or did you sit down, open Excel, and start figuring things out because you had a deadline?

i bet it was the second one. And you probably learned more in the first hour of actual use than you would have in a week of lectures about spreadsheet theory.

AI tools work the same way. You learn by doing. By trying. By failing. By asking the tool a question and getting a useless answer and then rephrasing the question and getting a better answer. By discovering that it's amazing at this task and absolutely terrible at that one. By developing an intuition for what it can handle and what it can't.

No course can give you this intuition. Only use can.

Here's a concrete example. I worked with a team of project managers who all completed the same AI training course. Six weeks of modules, quizzes, and a final assessment. They all passed. Then I asked them to use AI to help with their actual project management work.

The ones who hadn't touched an AI tool outside the course were lost. They knew what a prompt was in theory but couldn't write an effective one. They knew about hallucinations from the course materials but couldn't spot one in practice. The course had given them knowledge without competence.

The two people in the group who had been using ChatGPT on their own for a few months? They were miles ahead. They hadn't taken the course. They'd just been using the tool. And they were more capable than everyone who had the certificate but not the practice.

The courses that are actually worth something

i'm not saying all courses are worthless. Some have genuine value. But you need to know what to look for.

Good courses are hands-on. Every module should require you to actually use an AI tool. Not watch someone else use one. Use one yourself. If you can complete the course without opening ChatGPT or Claude, it's a theory course, not a skills course.

Good courses are specific to your domain. "AI for HR professionals" is more useful than "Introduction to AI." "Using AI for financial analysis" is more useful than "Machine Learning Fundamentals." The more specific to your actual work, the better.

Good courses are short. If a course takes more than 10 hours, it's probably padding. The core skills of using AI tools at work can be taught in a few hours. Everything beyond that is context and nuance that you'll learn better through practice.

Good courses admit their limitations. Any course that claims to make you an "AI expert" is lying. Any course that tells you to keep practising after you finish is being honest.

The best course i've ever seen for non-technical professionals was 3 hours long. It explained the basics, walked through practical examples, and then said "now go use this every day for a month and come back for the advanced session." The month of practice was the actual course. The videos were just the introduction.

What to do instead

Here's the alternative to taking a course. It takes less time and produces better results.

Day 1: Open an AI tool (ChatGPT, Claude, Gemini, whatever). Ask it to help you with a real work task. See what happens. Spend 30 minutes.

Days 2-5: Keep using it. Try different types of tasks. Drafting, analysing, summarising, brainstorming, formatting. Find what it's useful for in your specific role.

Week 2: Focus on the tasks where it helped most. Refine your approach. Try different ways of asking for things. Compare outputs. Start developing a feel for good prompts vs bad prompts.

Week 3: Build something. Take your best use case and turn it into a workflow or tool your team can use. Something concrete. Something demonstrable.

Week 4: Show someone. Your manager. A colleague. Your LinkedIn network. Share what you built and what you learned. Get feedback.

Four weeks. Zero courses. One month of daily practice. You'll be more capable than 90% of people who completed an AI certification. Not because certifications are bad, but because practice is better.

The certification question

"But won't employers want to see a certification?"

Some might. For most roles, no. What employers increasingly want to see is evidence of applied AI capability. "I used AI to reduce our report generation time by 60%" is worth more than "I have a certificate in AI Fundamentals from Coursera."

If you're in a regulated industry where specific certifications are required, obviously get those. But for most people in most roles, your time is better spent building a portfolio of AI projects than collecting certificates.

Think about what goes on your CV. Which sounds better?

Option A: "Completed 40-hour AI Fundamentals certification (2026)"

Option B: "Built AI-powered reporting workflow saving team 15 hours/week. Led department AI pilot programme. Trained 8 colleagues on practical AI tool usage."

Option B wins every time. And Option B doesn't require a single course. It requires practice.

The uncomfortable truth about why people choose courses

Courses feel safe. They have structure. They have clear outcomes. You watch the video. You do the quiz. You get the certificate. Progress is measurable and non-threatening.

Actually using AI tools is messy. You don't know what you're doing. The tool gives weird outputs. You make mistakes. You feel incompetent. It's uncomfortable.

But discomfort is where learning happens. The course removes the discomfort and, with it, most of the learning. The person who spent a frustrating afternoon trying to get ChatGPT to properly analyse their data learned more than the person who watched a polished video about data analysis with AI.

i'm not trying to shame anyone who's taking courses. The impulse to learn is good. The anxiety that drives it is understandable. I'm just trying to redirect the energy toward something that actually works.

AI literacy isn't what you think it is. It's not knowledge. It's capability. And capability comes from doing, not studying.

The one thing to do today: cancel the course you're thinking about signing up for. Instead, open an AI tool and spend 30 minutes using it for real work. Tomorrow, do it again. Do that every day for a month. At the end of the month, you'll know more than any course could teach you. That's not an exaggeration. That's just how skills work.