AI Scams and How to Spot Them

(And Why Helen Nearly Sent £2,000 to a Stranger)

Helen in Poundbury got a phone call last Tuesday afternoon.

It was her grandson's voice. She'd know it anywhere.

"Gran, I'm in trouble. I've been in an accident. The police are here. I need £2,000 for the solicitor right now or they're going to charge me. Please don't tell Mum, she'll be so upset. Can you transfer it to this account?"

He sounded panicked. Genuine. Desperate.

Helen's hand was on her laptop, about to log into her banking, when something made her pause.

"What's your sister's name?" she asked.

Silence. Then the line went dead.

Her grandson doesn't have a sister. Helen had just encountered a voice-cloning AI scam. And she very nearly fell for it.

This week's Tech Tuesday is different. It's not about how to use AI. It's about how criminals are using it against you. And what you can do to protect yourself and your family.

What AI Scams Actually Look Like

AI scams aren't science fiction anymore. They're happening right now, in Poundbury and everywhere else. The technology that helps you plan meals and understand your energy bill? Criminals have the same technology. And they're using it in increasingly sophisticated ways.

Here's what's actually happening out there.

Voice Cloning Scams (The Grandson Trick)

This is the one Helen encountered. It's becoming frighteningly common.

Criminals find a short video of someone speaking on social media. Maybe a birthday message. Maybe a TikTok. Thirty seconds is enough.

They feed that voice into AI software. The AI learns to mimic it perfectly.

Then they call elderly relatives pretending to be in trouble. Car accident. Arrested. Injured. Medical emergency. Always urgent. Always needing money immediately.

The voice sounds exactly like your grandson. The panic sounds real. Your instinct is to help.

An expert perspective:

This week we have another expert's perspective, from Tony Merrill, who spent over twenty years working in defence and security. Now retired in Poundbury with his partner Vivienne, Tony has seen how these scams evolve.

"The voice cloning technology is remarkable," Tony explains. "Five years ago, you needed hours of audio to clone a voice. Now, thirty seconds from social media is enough. The quality is so good that even close family members can't tell the difference."

How to protect yourself:

• Have a family code word that only you and your family know. If someone calls claiming to be family in trouble, ask for the code word.

• Ask questions only the real person would know. "What did we have for dinner last Sunday?" "What's your dog's name?" Simple questions criminals won't have answers to.

• Never transfer money based on a phone call alone. Always hang up and call the person back on their known number.

• Tell the caller you need to check with another family member first. Real family emergencies can wait five minutes. Scammers will pressure you to act immediately.

  
  

Remember: Urgency is always a red flag. Real emergencies rarely require instant bank transfers.

Deep Fake Video Calls (The CEO Fraud)

This one targets businesses, but it's worth knowing about.

Criminals use AI to create video calls where they appear to be your boss, your CEO, or a trusted colleague.

They join a Zoom call. Their face looks right. Their voice sounds right. They ask you to authorise a payment or send confidential information.

It's all fake. The entire video feed is AI-generated in real time.

"This technology used to be available only to major film studios," Tony says. "Now anyone can download the software for free. The barrier to entry for sophisticated fraud has essentially disappeared."

Small business owners in Poundbury need to be aware:

• Never authorise large payments based solely on a video call. Use additional verification.

• If something feels off, it probably is. Trust your instincts even if the video looks perfect.

• Implement verification procedures. If your boss asks for an unusual payment, confirm through a different communication channel.

• Be suspicious of last-minute urgent requests, especially involving money or confidential data.

• AI-Written Phishing Emails (The Grammar Trick Doesn't Work Anymore)

Phishing emails

Remember when you could spot a scam email by the terrible grammar and spelling mistakes?

Those days are over.

Criminals now use ChatGPT and other AI tools to write perfect English. The emails are grammatically flawless, professionally formatted, and completely convincing.

"The old advice was to look for spelling errors," Tony notes. "That advice is now obsolete. AI writes better English than most native speakers."

Recent AI-generated phishing emails look like:

• Convincing messages from your bank about suspicious activity

• Professional-looking invoices from companies you actually use

• Realistic job offers with detailed descriptions

• Genuine-seeming messages from HMRC about tax refunds

• Authentic-looking delivery notifications from Royal Mail or courier services

  
  

How to spot them now:

• Check the sender's email address carefully. It might say "Barclays" but the address is barclays-security-uk.com.ru- notice the .ru at the end.

• Hover over links without clicking. The display text might say lloyds.co.uk but the actual link goes somewhere else entirely.

• Don't click links in unexpected emails. Go directly to the website by typing the address yourself.

• Be suspicious of unexpected attachments. Don't open them. Contact the supposed sender through official channels to verify.

  
  

Remember: Your bank will never ask for your full password or PIN by email. Neither will HMRC. Neither will legitimate companies.

Fake Customer Service Chatbots

You have a problem with your broadband. You search "BT customer service contact."

The first result looks official. You click it. There's a chat window. You explain your problem.

The chatbot asks you to verify your account. You provide your account number, maybe your postcode, perhaps your mother's maiden name for "security."

It's all fake. The website is fake. The chatbot is AI. You've just given criminals your personal information.

"The scam websites often rank higher than the real ones in search results," Tony warns. "Criminals pay for advertising to appear at the top of Google searches. People assume the top result is legitimate."

How to protect yourself:

• Never click on ads in search results. Scroll down to the actual organic results, or better yet, type the company's address directly.

• Bookmark the real customer service pages of companies you use regularly. Use your bookmarks, not search engines.

• Be suspicious of chatbots asking for sensitive information. Real customer service rarely needs your full account details in a chat.

• If something feels wrong, close the window and call the official customer service number from the company's official website.

• Verify you're on the right website. Check the address bar. Is it actually bt.com or is it bt-help-uk.com?

Romance Scams with AI-Generated Photos

This one's heartbreaking and increasingly common.

Someone seems perfect on a dating site. Their photos are attractive but not supermodel-level. They seem genuine.

You chat for weeks. They're charming, interested, understanding. A real connection develops.

Then they have an emergency. Medical bill. Stranded abroad. Investment opportunity. They need money. Just temporarily. They'll pay you back.

The person doesn't exist. The photos are AI-generated. The personality is scripted. Your money is gone.

"AI image generation has made this frighteningly easy," Tony explains. "Criminals can create hundreds of realistic-looking profiles in minutes. Each with consistent photos of the same fake person at different ages, in different settings."

Warning signs:

• They never want to video call. Always an excuse. Bad camera. Shy. Broken phone. No exceptions ever.

• Photos look professional but generic. AI-generated faces often have a certain smoothness that's hard to describe but noticeable.

• They move very quickly from the dating site to private email or messaging. They want to get away from the platform's safety features.

• Money requests always come with elaborate stories. Emergency. Stranded. Investment. Never straightforward.

• They claim to be local but are currently abroad. Military deployment. Working overseas. Visiting family. Always a reason they can't meet in person.

  
  

What Tony Wants You to Remember

I asked Tony what the single most important thing people should know about AI scams.

"The technology is ahead of public awareness," he said. "Most people still think of scammers as criminals in far-away countries sending badly-written emails. But modern scams use AI to sound like your grandson, look like your boss, and write like a professional."

"The good news is the human element is still there. Scammers still rely on urgency, pressure, and emotional manipulation. Those warning signs haven't changed."

His key advice:

• Trust your instincts. If something feels wrong, it probably is. Even if you can't articulate why.

• Slow down. Criminals need you to act quickly before you think. Take your time. Real emergencies can wait five minutes.

• Verify through different channels. If someone calls, hang up and call them back. If someone emails, phone them directly. Never use contact details provided in suspicious messages.

• Tell someone else. Scammers want you isolated and panicked. Talk to a friend or family member before taking action.

• Remember that urgency is always a red flag. Legitimate organizations give you time to think.

Protecting Your Family

The people most vulnerable to these scams aren't stupid or gullible. They're trusting, kind, and want to help people they care about.

That's why it's important to have conversations now, before anything happens.

Talk to your family about:

• Setting up a code word for emergency situations. Something simple everyone can remember.

• Agreeing never to send money based on a phone call or message alone. Always verify through another method.

• Knowing it's okay to say no or ask questions. Even in apparent emergencies.

• Understanding that real banks and government agencies never ask for full passwords or PINs.

• Checking in with each other if something seems suspicious. No judgment. Just support.

What Helen Did Next

Helen hung up on the fake grandson and immediately called her real grandson.

He was fine. At work. Completely safe. No accident. No arrest. No emergency.

"I felt so stupid," Helen told me. "Then I got angry. How dare they use my love for my family to try to steal from me?"

Helen did something smart. She told everyone she knows about the call. She's sharing her near-miss as a warning.

"If my story stops one other person from falling for this, then something good came from it," she said.

That's why she agreed to be featured in this article.

The Bottom Line

AI is a tool. Like any tool, it can be used for good or bad.

We've spent weeks learning how to use it helpfully. Plan meals. Understand bills. Write emails. Learn about the world.

Now you know how criminals are using the same technology to steal from you.

The warning signs haven't changed: Urgency. Pressure. Requests for money. Too good to be true.

What has changed is how convincing the scams can be. Perfect grammar. Cloned voices. Real-looking faces. Professional formatting.

Your best defence is the same as it's always been:

Slow down. Think. Verify. Talk to someone you trust.

And remember Helen's question: "What's your sister's name?"

One simple question saved her £2,000. One question you know the answer to can save you too.

Stay safe. Stay sceptical. And look after each other.

Next week: Back to the positive side - AI Tools for Gardening in Summer

Got a Tech Tuesday question or suggestion? Email tech@lovepoundbury.org

Special thanks to Tony Merrill for contributing his security expertise to this article.