Can you spot an AI scam?

For this week’s Reality Cheque, we’re looking at the money myth that AI is easy to spot, and so are AI scams.

You’re practised at ignoring texts from a sus “CRA” number, offering imaginary tax refunds. And emails from Nigerian princes go straight into the trash. But are you ready for the next generation of AI-powered fraud?

“There are a lot more scams and they’re getting more sophisticated,” says KPMG fraud specialist Marilyn Abate. “It’s much more real now.”

Not only are the fraudsters using AI to generate scams, making it much faster and easier to tailor, create and send them, but AI tools themselves can be a risk, Abate says. “I would never put my personal information into ChatGPT or any of these other online chats,” she says. “You really don’t know where it ends up.”

There are other ways to both prevent and avoid AI scams, too. Here are some things to think about.

AI and fraud: What’s really happening?

Scams are everywhere. In Edmonton, for instance, there were more than 4,000 reported frauds in 2025, with losses of more than $58 million. About $25 million of that was from investment and crypto scams; other contributors included identity theft, phishing scams and compromised credit and debit accounts. In total, according to the Canadian Anti-Fraud Centre (CAFC), Canadians reported more than $704 million in fraud losses in 2025.

While not all were AI-related, AI is a major contributor to scams becoming more and more sophisticated and even harder to detect. And Canadians are concerned: a 2026 survey by Interac found that 79% believe scams are more convincing and difficult to spot than ever, thanks to AI. That means “many of the traditional red flags Canadians once relied on are no longer as reliable,” says David Chin, leader of fraud strategy at Interac. “One in two Canadians are increasingly second-guessing even legitimate messages.”

How AI scams work

Those AI tools you use to work more efficiently also help fraudsters, allowing them to more quickly tap into current events and produce messages aimed at specific audiences, like the tariff-related payment request scams that 58% of Canadians have received in the past six months. These tools are also helping scammers make their messages believable, says Chin: “Scams no longer look like scams [of the past]. Today’s scams are polished, contextual and often tailored to what’s happening in the news in your own life.”

There’s also the flip side of AI-enabled scams: those that take the information you give to AI tools and turn it into bespoke fraud, whether it’s attempts at identity theft made easier by hacking into your personal information or fake bank ads popping up after you’ve been researching mortgages. 

“All of a sudden, that link is not the real bank’s, it’s a fake website,” says Abate—you might click and think you’re applying for a mortgage with a known bank, when you’re really giving your personal information to a fraudster. In addition, Chin points out, “scammers are now impersonating AI platforms themselves.” It’s not just fake websites, emails or text messages, but full tools that are fraud.

And then there’s deepfakes: AI-produced images or recordings impersonating celebrities, politicians or even people you know personally. In the fraud world, they’re designed to convince you to cough up money, whether it’s for a “surefire” crypto investment deal or to bail your grandchild out of a sticky situation. “Those are things you’ve got to be careful of,” Abate says. “You have to analyze it with your mind, not your heart.”

How to safeguard yourself from AI scams

In our hectic world, it’s easy to click quickly and make a move you’ll regret. But the number-one guideline to avoid scams is to pause, consider and verify. “Scams are designed to pressure people into acting quickly, often before they’ve had time to think,” says Chin. “Pausing to scrutinize can disrupt that momentum and help prevent costly mistakes.”

And whereas telltale signs of fakery have typically included off-colour logos, misspelling and strange formatting, that’s not necessarily the case anymore. “Traditional warning signs are no longer reliable indicators of fraud,” Chin says. So, even if a message comes from a trusted organization or person, verify it by contacting them directly. For example, rather than clicking a link, you can open an institution’s website or app yourself, without using the link, or call the number on your credit or bank card.

Tips on avoiding AI fraud

Hopefully, you’re already using two-step verification for logins. And it’s best to keep personal details to yourself rather than sharing them on Facebook or with a chatbot—think your full name, address, phone number, birthplace, first pet’s name and certainly PINs and two-step verification codes. “Anything you enter into an AI tool should be treated as information you wouldn’t mind being public,” Chin says. (Here’s how a former hacker admits to stealing passwords.)

Discuss AI scams with friends and family members, too, especially those who might not keep up to date on tech trends. Abate says that some families and even companies are coming up with safe words to ensure the person you’re talking to is who they say they are. 

Plus, there’s that old adage: if it seems too good to be true, it probably is. And if you need to act fast, do the opposite.

By pausing to think instead of simply reacting, your spidey sense will kick in and help you spot fraud. “Ultimately, trust your instincts,” says Chin. “If a message feels off, whether because of the timing, the tone or the pressure to act, that’s a strong signal to stop, verify and seek out trusted sources before taking any action.”