Welcome! Let's start with a question
Have you ever had a message that felt a bit off, but you couldn't quite say why? Maybe an email that looked official, or a text that seemed to come from someone you know?
If so, you're not alone. And here's the thing we'll explore together in this first level. Scams haven't just changed a little in the last few years. They've had a big upgrade, and the reason is artificial intelligence, or AI.
Don't worry. You don't need to be a tech expert to stay safe. By the end of this lesson you'll have one simple habit that beats almost every trick. Let's go.
A quick story: Margaret's phone call
Let's meet Margaret. She's 72, and one evening her phone rings. It's her grandson, Tom. Or at least, it sounds exactly like Tom.
'Gran, I'm in trouble. I've had a car accident and I need £800 for a solicitor tonight. Please don't tell Mum, she'll be so angry. Can you send it now?'
His voice is shaky. It's his laugh, his way of saying 'Gran'. Margaret's heart races. She reaches for her bank card.
But Tom was at home the whole time, watching football. The voice was fake, made by a computer.
So how did a scammer pull that off? Let's find out.
Three ways AI helps scammers
Think of AI as a super-fast, very cheap assistant. Scammers used to do everything themselves, slowly and often badly. Now they have a helper that never gets tired. Here are three big things it does for them.
1. Perfect writing. Do you remember when scam emails were full of spelling mistakes and odd grammar? Those mistakes were a handy warning sign. Now AI can write a polished, friendly message in seconds, in any language and any tone. It can even write thousands of slightly different versions, each one tailored to a different person.
2. Cloned voices. This is what happened to Margaret. AI can copy someone's voice from a short clip of audio, like a video posted on social media. Starling Bank has warned that just three seconds of audio can be enough. The fake voice can then say anything the scammer types.
3. Fake faces. AI can create realistic photos and videos of people who don't exist, or make real people appear to say things they never said. These are often called 'deepfakes'. A scammer can invent an attractive new 'friend' online, or show a famous face appearing to back a money-making scheme.
Why this makes scams more common, not just smarter
Here's the part people often miss. AI doesn't only make scams more convincing. It makes them cheap and quick to produce.
Imagine a forger who once spent a week painting one fake picture. Now picture a machine that prints a thousand a minute. That's the difference AI makes.
And the numbers back this up. UK Finance's 2026 fraud report found that nearly £1.3 billion was stolen by fraudsters in the UK in 2025. There were more than four million cases where money was taken. That's almost eight every minute. The report names AI as a major reason fraud is growing in scale and sophistication.
So it's not a question of whether you'll see a scam. You almost certainly will. The question is whether you'll spot it.
The good news: the tricks underneath haven't changed
Now, this is where we can breathe a sigh of relief. AI changes the costume, but not the person inside it.
Almost every scam still pulls the same emotional levers. Look back at Margaret's call. Can you spot them?
- Hurry: 'Send it now, tonight.'
- Fear: an accident, trouble with the law.
- Secrecy: 'Please don't tell Mum.'
- Reward: not in Tom's call, but common elsewhere, like 'You've won a prize' or 'Double your money'.
Why do these work? Because when we feel rushed, scared or excited, our thinking brain takes a back seat. Scammers know that. They want you acting, not thinking.
So here's a useful rule. However real the voice, face or message seems, if you feel hurry, fear, secrecy or a too-good reward, treat that feeling as an alarm bell.
Your new habit: pause, then check another way
So what should Margaret have done? It's simpler than you might think.
First, pause. No real emergency falls apart because you took five minutes. A real grandson will understand.
Then, check through a channel you already trust. That means hanging up and contacting the person or organisation yourself, using a number or address you already know. That could be Tom's number saved in her phone, or calling his mum. Not the number that just rang, and not a link in the message.
This is the same idea behind the UK's Take Five to Stop Fraud campaign: stop, then challenge whether the request could be fake.
Why does this work so well? Because a scammer can fake a voice or a face, but they can't easily answer Tom's real phone. When you step outside the conversation they've set up, their whole trick collapses.
And if it turns out to be real? You've lost a few minutes. That's a very good trade.
Key takeaways
- AI helps scammers in three big ways: perfect writing, cloned voices and fake faces.
- Because AI is cheap and fast, scams are now far more common, not just more convincing.
- Almost every scam uses the same pressure tactics: hurry, fear, secrecy and a promise of reward.
- When you feel that pressure, pause.
- Then check through a channel you already trust, like a number you already have, never the one the message gives you.