What is AI, and how will it change our lives?

A phone call to a cheap hotel

In the early hours of 8 October 2024, a phone rang in a hotel room in California.

The man who picked up was 76 years old. He was a professor from Toronto. The call was from Stockholm. He had just won the Nobel Prize in Physics.

His name is Geoffrey Hinton. He said he was "flabbergasted".¹ On the phone to reporters in Stockholm, he explained he was in a cheap hotel with a poor internet and phone connection. He had planned an MRI scan that day. He joked he would have to cancel it.²

Here is a detail worth knowing. Hinton once said he "dropped out of physics after my first year in university because I couldn't do the complicated math."¹ Decades later, physics gave him its highest prize.

He won it for work on artificial neural networks, the idea that a computer can learn a little like a brain does.¹ He has taught at the University of Toronto since 1987, and much of his later work happened there, in the city I call home.

The next day, it happened again. Half of the Nobel Prize in Chemistry went to two people at Google DeepMind, Demis Hassabis and John Jumper. Their AI model, AlphaFold, solved what the Academy called "a 50-year-old problem": predicting the shape of proteins. The other half went to David Baker, who designs new proteins with computers.³

Two Nobel Prizes. Two days. Both rooted in AI.

That week, something quietly changed. AI stopped being a clever tool that scientists used. It became the science.

But here is the funny thing. Many smart people, people who run businesses and teams, still feel a little lost when someone says "AI". They use it. They read about it. But if you asked them to explain it to a 15-year-old, they would go quiet.

If that's you, you're in good company. This article is for you.

By the end, you'll know what AI is, how the tools you use actually work, where AI already helps us, where it goes wrong, and what I think you should do about it. No maths. No jargon. I promise.

So what is AI, really?

Let's start with a definition that governments around the world now use. It comes from the OECD, and it's a little dry, so I'll explain it after:

"An AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments."⁴

In plain words: AI is a computer system that looks at what you give it and works out a useful answer. The answer can be a prediction, a piece of writing, a suggestion or a decision.

The key word is "infers". A normal computer program follows rules a person wrote. AI works out its own rules from examples.

Think of a spam filter. Nobody can write a rule for every spam email. So we show the computer millions of emails marked "spam" or "not spam", and it learns the patterns. That's machine learning.

AI is a family of ideas, and each one sits inside the one before. Tap each ring to see what it means.

Is it all AI? Not quite. A calculator is not AI. Neither is a spreadsheet formula. They follow fixed rules and give the same answer every time. Try this quick quiz.

How did we get here?

AI is not new. The question is older than most of our parents.

In 1950, the British mathematician Alan Turing wrote a paper that began: "I propose to consider the question, 'Can machines think?'"⁵

Five years later, four American scientists wrote a short proposal for a summer workshop at Dartmouth College. They wanted "a 2 month, 10 man study of artificial intelligence." Their big idea was that "every aspect of learning or any other feature of intelligence can in principle be so precisely described that a machine can be made to simulate it."⁶

That proposal gave the field its name. They hoped to make real progress in one summer. It took about seventy years.

Why did it take so long? Three things had to arrive together.

1. Data. The internet gave us billions of pages, photos and conversations to learn from.
2. Computing power. Chips built for video games turned out to be perfect for training AI.
3. Better ideas. Researchers like Hinton kept working on neural networks for decades, even when most people thought they were a dead end.

When all three came together, around 2012, progress stopped crawling and started running.

How ChatGPT actually works

This is the part most people never get told, and it changes how you use these tools.

A large language model (LLM) like ChatGPT or Claude does one thing: it predicts the next word. Then the next. Then the next.

It learned to do this by reading a huge amount of text. Along the way it picked up grammar, facts, styles and a lot of what we'd call knowledge. But it is not looking things up in a database. It is guessing what words are most likely to come next, very, very well.

Try it below. The model sees the start of a sentence and weighs its options.

Look at that last example. When the model writes a number, it is still predicting. Three different totals look almost equally likely. It will pick one, and it will sound completely sure.

That's why AI can be confidently wrong. The industry calls it "hallucination". I call it a very good reason to check the totals.

Hinton put it beautifully in his Nobel banquet speech. This new kind of AI, he said, "excels at modeling human intuition rather than human reasoning."⁷ It has a great gut feeling. It is not always careful.

AI is already in your life

You don't need to wait for the future. It's already here.

When ChatGPT launched in late 2022, analysts estimated it reached 100 million monthly users in about two months, faster than any app before it.⁸ By February 2026, OpenAI said 900 million people were using it every week.⁹ Stanford's AI Index found that generative AI reached 53% adoption in just three years, faster than computers or the internet.¹⁰

Businesses are following, here in Canada too. Statistics Canada asks businesses each quarter if they used AI to make goods or deliver services. Look at how fast the number has grown:

It roughly tripled in two years.¹¹ And it isn't even across industries. Finance and insurance is at 40.4%. Professional services, which includes accounting, is at 32.4%. Construction is at 9.2%.¹¹

One more number stands out. 40% of Canadian businesses said AI was "not relevant" to them.¹¹ I understand why. But I'd gently ask: is that still true, or is it just not urgent yet?

Big companies are almost all in. McKinsey's 2026 survey found nearly nine in ten organisations use AI somewhere. But only 37% could point to any effect on their profit.¹² So using AI and getting value from it are two different things. We'll come back to that in later articles.

How AI will change our lives

Let's look at where it already makes a real difference. I'll stick to what has been measured, not what has been promised.

Science and health

Remember AlphaFold. Before it, working out the shape of a single protein often took "a year or more of expensive, painstaking experimental work."¹³ AlphaFold has now predicted more than 200 million of them. More than three million researchers in over 190 countries use it.¹³

Why does that matter to you? Proteins are the machinery of life. Knowing their shape helps scientists design medicines and understand diseases. A starting point that once took a year in the lab can now be looked up.

In hospitals, a large trial in Sweden, with more than 100,000 women, tested AI-supported breast cancer screening. It found 29% more cancers, while cutting the radiologists' reading workload by 44%.¹⁴ That's more lives caught early, with less strain on busy doctors.

Work

This is the question most people ask. Will AI take my job?

The International Monetary Fund looked at this carefully. It found that almost 40% of jobs around the world are "exposed" to AI, and about 60% in advanced economies like Canada.¹⁵

"Exposed" doesn't mean "replaced". Statistics Canada split it further. About 31% of Canadian workers are in jobs that are highly exposed and where AI could do much of the work. But another 29% are highly exposed in a way where AI is more likely to help them than replace them.¹⁶

The World Economic Forum asked more than 1,000 large employers about the rest of this decade. They expect 170 million new jobs and 92 million lost by 2030, a net gain of 78 million. But they also expect almost 40% of the skills we use at work to change.¹⁷

So the jobs don't simply disappear. They change shape. And the people who learn the new shape do well.

There's good news in the research too. A study of more than 5,000 customer support agents found that an AI assistant helped them solve 14% more issues per hour. For new and less experienced agents, the gain was 34%.¹⁸ AI helped the newest people the most. It shared the know-how of the best people with everyone.

That gives me hope. Used well, AI can be a great equaliser.

Money and the economy

Economists are still arguing about the size of the prize. McKinsey estimated generative AI could add $2.6 to $4.4 trillion to the world economy each year.¹⁹ Goldman Sachs estimated a 7% lift to global GDP over ten years.²⁰

I'd treat those as direction, not destination. Forecasts are opinions. But the direction is clear.

Where it goes wrong

I'd be doing you a disservice if I only told you the good part. AI is fast at the routine. It is risky when we trust it blindly.

The chatbot that made a promise

In November 2022, Jake Moffatt lost his grandmother. The same day, he asked Air Canada's website chatbot about bereavement fares.

The chatbot told him he could book now and claim the lower bereavement fare within 90 days. He trusted it and bought tickets between Vancouver and Toronto. But the airline's own policy page said the bereavement fare could not be claimed after travel. Air Canada refused the refund.

He took it to British Columbia's Civil Resolution Tribunal. Air Canada argued, in effect, that the chatbot was a separate legal entity responsible for its own actions.

The tribunal member was not impressed: "It should be obvious to Air Canada that it is responsible for all the information on its website."²¹ Air Canada had to pay him $812.02.

A small amount of money. A big lesson for every business owner: you own what your AI says.

The lawyers and the fake cases

In New York in 2023, two lawyers filed a court brief that cited six court cases. None of them existed. ChatGPT had made them up, complete with names and quotes. The judge fined them $5,000.²²

The AI wasn't lying. It was predicting what a court case citation should look like. Remember the next word game.

The CFO who wasn't there

In early 2024, a finance employee at the engineering firm Arup in Hong Kong joined a video call with the company's UK-based CFO and several colleagues. He was asked to make some confidential transfers. He made 15 of them, worth about US$25 million.

Everyone else on the call was a deepfake.²³

As a finance person, this is the one I'd show every team. The answer is not more fear. It's better controls: call back on a known number, two people approve, and no payment on a video call alone.

The costs we don't see

AI also uses a lot of electricity. The International Energy Agency expects data centres worldwide to use about 945 terawatt hours by 2030, slightly more than all of Japan uses today.²⁴

And there's fairness. The IMF warned that "in most scenarios, AI will likely worsen overall inequality."¹⁵ That isn't a reason to stay away from AI. It's a reason to make sure more people, and more small businesses, get to use it well.

Who sets the rules?

Governments are catching up, at different speeds.

The European Union passed the first broad AI law in 2024 and is bringing it in step by step. In Canada, an earlier federal AI bill died in January 2025 when Parliament was prorogued. Canada then appointed its first Minister of Artificial Intelligence and Digital Innovation, and published a new national AI strategy in 2026, but there is still no single AI law.²⁵²⁶ The United States has, so far, no broad federal AI law and has favoured a lighter touch.

For most businesses, the practical message is simple. The law may be unclear, but your responsibility isn't. If your AI harms a customer, the Air Canada case shows who pays.

My take

When I started this series, I wrote that the world is going through a transformation it has never seen, and that the sooner we're prepared, the better. I still believe that.

But I also believe something I wrote years ago, long before ChatGPT: we are not machines, far more than that.

AI is very good at the routine. It's getting good at the complex. But judgement, care, trust and responsibility still belong to people. A person signs off. Never the tool.

And here's what matters most to me. Good financial advice has mostly been something only bigger companies could afford. I believe AI can change that. It can bring good advice to every business, not just the big ones. I want to help make that happen.

The next five years

In this day and age, it's almost impossible to predict what comes next week. The technology is changing that fast. It took us decades to get here. But a lot can change in the next five years.

Here's my view. In five years, we're still humans. New industries and new jobs will be created. The line between AI and humans will get thinner. Businesses will go through big changes, and many will employ fewer people. A handful of companies, individuals and governments will become more powerful, with more control.

And the people and businesses that use AI well will do much better than they do now. Those that don't will fall behind.

Five things you can do this week

You don't need a course or a budget to start. Tick these off as you go.

What comes next

This is the first of seven articles I'll publish over the next seven months, one each month. Next month, I'll look at why AI adoption is no longer a choice, and why starting small, now, beats starting big, later.

Then we'll go deeper: what AI means for small businesses, for the finance function in startups and growing companies, for finance professionals and their careers, and how anyone can start building their own AI agents. In the last one, I'll share where I think all of this is taking us.

I'm writing these with the help of AI, and I'm learning as I go. There could be areas I miss. So I'd really value your thoughts.

What is one question about AI you've never asked out loud?