Spot the Scam: Every day, people across Ottawa and Canada are falling victim to scams. This week, CTVNewsOttawa.ca will share their stories and advice on how to spot a scam. Fraudsters are increasingly turning to easily available AI tools, such as video and voice generating programs, to execute scams in Canada. According to EY Canada’s 2025 Cybersecurity Threat Outlook, AI-generated deepfakes comprised nearly 25 per cent of fraud attempts against biometric authentication last year. “Throughout 2024, we saw deepfake technology drive increasingly sophisticated scams. In these scenarios, criminals clone voices to impersonate CEOs, loved ones and others — duping people into losing significant amounts of money. In fact, Canadians lost the greatest amount, per person, among those surveyed globally,” the report says. “Threat actors are launching a number of scams around Canada in the space of deepfakes where they have the ability to call someone with someone else’s voice, the ability to call someone on a video call with someone else’s face and voice, but also combine that with a sense of urgency, loss, or something of that nature that makes people fall for it,” said Umang Handa, a cybersecurity expert and partner at EY Canada, speaking to CTV News Ottawa. The EY Canada report says 27 per cent of Canadians reported encounters with deepfake scam calls. Last year scams cost Canadians an average of $1,479. “Beyond voice fraud, attackers are also employing generative AI for persuasive phishing emails, malicious code and business email compromise,” the report says. Handa says a year or two ago, it was easier to spot deepfakes because of glitches and inconsistencies in the video or sound, but the technology is getting better. “There are professional tools that have the ability to do it and they’re available for free for the most part,” he said. “The sense of realism has blurred. Whatever you see, whatever you hear, it’s very real. You only need about two minutes of someone’s clip, of someone’s voice to create a deepfake of it.” Handa says that while scams are becoming more sophisticated and harder to spot, the good guys are also working hard to stop them. “The losses are there. There’s always going to be a place where the bad guys are misusing the technology when it comes out. Having said that, the coalition we’ve formed in Canada of the banks, the telcos, media consulting organizations, cybersecurity and others, they’re trying to help,” he said. “It still starts with education. There are a lot of educational campaigns being launched to talk about deepfakes, to talk about frauds and now there’s technology that are coming up that will be able to tell us if it’s an AI or if it’s a person. While it’s not released at scale, I can see this coming out in the next few years as well.” The EY Canada report says regulatory changes in Canada and abroad, such as the Critical Cyber Systems Protection Act and the EU Artificial Intelligence Act will also make a difference. Can you use AI to spot an AI scam? Handa says tools like ChatGPT and AI detectors aren’t reliable when it comes to trying to determine if something else was made by AI. “I actually used ChatGPT to create an AI image of somebody, completely fake, and I put it into another tool like that and the other tool was not able to tell me if it this was an AI-generated image,” he said. “To some extent, you can use that to help you out, but you can’t rely on the results. But that definitely is one of the tools. In fact, AI in itself is being used to determine if something itself is AI, but it’s not foolproof.”
