Fraude et usurpation d'identité facilitées par l'IA

Generative AI has made convincing impersonation — by voice, video, and text — accessible to fraudsters with no specialized technical expertise. This structural shift has produced significant financial harm to Canadians.

In March 2023, eight seniors in St. John's, Newfoundland lost $200,000 in three days to a grandparent scam ring that used suspected AI voice cloning to impersonate family members in distress. CBC Marketplace's 2025 investigation confirmed that current voice cloning tools can produce convincing replicas from short audio samples — a phone greeting or social media video is sufficient.

At a larger scale, AI-generated deepfake videos of celebrities and public figures — including Elon Musk, Dragon's Den personalities, and Prime Minister Mark Carney — were used in cryptocurrency investment scams, particularly during and after the 2025 federal election. Over 40 Facebook pages ran fraudulent Carney deepfake schemes. Individual losses from deepfake-driven crypto scams were substantial: an Ontario woman lost $1.7 million in retirement savings to a scheme using a deepfake of Elon Musk; a Prince Edward Island man lost $600,000 in life savings to a similar scam. The Canadian Anti-Fraud Centre reported $103 million lost to crypto scams in 2025, with AI deepfakes as a major vector.

The structural condition is the asymmetry between fraud capability and detection capability. AI tools have made convincing impersonation cheap and scalable. Law enforcement forensic capacity, financial institution identity verification, and consumer awareness have not adapted. The fraud techniques that previously required criminal organizations with resources and expertise are now accessible to anyone with a laptop and free AI tools.