AI Is Changing Our Jobs. How Do We Learn What Comes Next?

If you wanted to learn basketball, you could read a book. You could watch games and listen to experts talk about shooting.

Sooner or later, though, you would need to pick up a ball.

That’s how I think about learning AI.

I learn better when I can watch someone work through a task. I want to see where they start, what decisions they make and what happens when something goes wrong. Then I want to try it myself.

Watching helps me get started. Practice helps me learn.

I remember being the beginner

Over the past two years, I’ve spent a great deal of time working with AI. Sometimes I have to think back to when I started.

I didn’t know what all the terms meant. I felt hesitant about jumping into something unfamiliar. It was like standing beside a pool without knowing how deep it was.

I try to remember that feeling when I help someone else. When you use these tools every day, it’s easy to forget how confusing the first few steps can feel.

That doesn’t mean someone can’t learn. They need a starting point they understand. For a business owner, that might be a customer enquiry, an estimate or a weekly report. Start with work they know, then show where AI can help and where it needs checking.

The worry behind the question

Someone in an operations role recently said to me, “Well thank goodness we don’t have AI yet because I think I’d be replaced.”

That comment stayed with me. What happens when a tool becomes better at parts of the work that have made you valuable?

In my earlier article, Will AI Take My Job?, I looked at that broader question. Here, I want to focus on what we can do next.

There’s no honest way to promise that everyone’s role will stay safe. There’s also no sound basis for saying every job touched by AI will disappear. The International Labour Organization estimates that one in four workers worldwide is in an occupation with some exposure to generative AI. Its assessment is that jobs are more likely to change than disappear entirely. Exposure isn’t the same as a count of future layoffs. ILO research

My view is that waiting for a perfect forecast leaves us with less time to learn.

Ask what an expert actually does

Someone I know was watching a panel about AI. The speakers were senior people, and she wondered: how much do they actually use it?

I think that’s a useful question. It doesn’t tell us whether those particular speakers were experienced. It reminds us to look beyond someone’s title.

People can have valuable experience in AI policy, research or strategy. Others spend their days building and testing things. Those are different kinds of experience.

If I want help completing a real task, I want someone who can show me how they do it. Show me the starting point. Show me the choices. Show me what failed and how you checked the result.

A polished demonstration can be helpful. Seeing the messy parts of a project can teach us even more.

A coach needs coaches too

This is part of why I want to coach people using AI. I like working through a real task with someone, from the first question to the finished piece. There are many decisions along the way that a list of prompts won’t explain.

But I need coaches too. When I enter an area I don’t know, I want to work with someone who has been there before. I don’t want to make an avoidable mistake just to prove I can work alone.

A good coach helps you build your own judgement. They can explain why they chose a tool, when they stopped trusting an answer and what made the final result ready to use.

Think of The Karate Kid. The useful connection is practice with feedback. You still have to do the work.

There’s more than one kind of AI expertise

Someone who has spent thousands of hours creating AI videos will have different experience from someone who has spent thousands of hours using AI for research. The same applies to programming, voice, writing, marketing and connecting business systems.

Hours alone don’t prove expertise. The work someone can show you matters. Ask whether they’ve done the kind of work you need, how they check quality and protect private information, and whether you’ll understand and manage the result afterwards.

Let AI help you learn, too

There’s a risk in having AI do everything before you understand any of it. You may get a finished answer without learning how to judge whether it’s good.

One study of high-school maths students found that unrestricted AI help improved practice performance but hurt later performance without the tool. A tutor designed to offer guided help largely avoided that harm. That’s one educational setting, not proof about every workplace, but the distinction is useful. Research in PNAS

When learning a new task, ask AI to explain its choices, offer a hint or review your attempt. Then explain the result back in your own words. If you can’t explain it, that tells you where you still need help.

What about trades and robots?

People often bring up plumbing, electrical work and other hands-on jobs. Those jobs can involve awkward spaces, changing conditions, safety responsibilities and problems that don’t follow a tidy script.

Statistics Canada’s research finds lower current AI exposure in many physical trades. That gives us context, not a promise of permanent protection. Trades research

Robotics adds another layer of uncertainty. A robot completing a demonstration doesn’t establish that it can safely and affordably handle a whole day of varied customer jobs.

Meanwhile, a trades business can examine enquiries, scheduling, estimates, records and follow-ups. Start where the business is losing time or disappointing customers. AI for trades and home services

Start with one real task

For an owner or founder, I’d keep the first learning project small:

1. Choose a familiar task you can review safely.
2. Watch someone do it and explain their decisions.
3. Try a fresh example yourself.
4. Check for mistakes, missing information and time spent fixing it.
5. Repeat, and write down where a person must remain involved.

If you have employees, give them room to practise too. A finished AI answer shouldn’t become the only training a junior employee receives. The goal is to build people who can use the tools and understand the work.

You don’t have to know everything first

I still find this field exciting. I also understand why it can feel overwhelming.

You don’t need to master every AI tool. Start with one useful task, learn from someone with relevant experience and practise until you understand what good work looks like. That’s the kind of practical AI coaching I want to provide through Salient AI.

We can’t control every change coming to our jobs. We can take part in learning how the work is changing.

At some point, we have to pick up the ball.

Written from Stephen Lau’s experience and point of view, with AI-assisted research, editorial production and illustrations. Published with Stephen Lau’s approval.