Startup deepidv doubles deepfake detection options and forms research partnership

Deepfake detection teams at Canadian startup deepidv and fraud prevention developer Scam.AI will work together under a new partnership that also includes software integration.

The partners will share datasets and perform joint research into deepfake detection. They say in a joint announcement the result is greater access to the latest synthetic media for faster detection updates.

Scam.AI’s deepfake detection will be offered as a premium module within the deepeye deepfake detection plugin for Chrome and its deepidv deepfake detection API. This decision allows deepidv customers to run either partner’s deepfake detection or layer them together.

A video posted to LinkedIn depicts a user responding to multiple prompts including repeating a phrase while taking a familiar selfie biometric video. The companies then perform “dozens of checks,” such as for boundary artifacts, lighting consistency and skin texture, to detect deepfakes in Android, iOS or WhatsApp.

Scam.AI’s deepfake detection in 98.2 percent accurate, according to the video.

The video also says the capability can be integrated with “the world’s best verification companies,” among which it lists Jumio, Sumsub, Veriff, Persona, Trulioo, Smile ID and Entrust.

Scam.AI revealed a strategic partnership with Qualcomm last month to produce on on-device deepfake detection model for live video calls on desktop computers.

Deepidv raised a million dollars in seed funding, opened an office in San Francisco and introduced a new suite of fraud detection tools in a busy March.

Deepening industry engagement

The same month, deepidv joined the Digital Identity Council of Canada (DIACC) to coordinate with the identity sector in its home country. Founder and CEO Shawn-Marc Melo called DIACC membership “a natural next step” for the company because of close mission alignment.

“Identity verification can no longer stop at document checks and basic selfie matching,” he says. “When deepfakes can fool liveness systems and synthetic identities can survive credit checks for years before busting out, the entire trust model needs to be rebuilt around continuous, forensic-grade verification. That is what deepidv delivers, and DIACC is the right coalition to help scale that standard across Canada and into global markets.”

The company released its “Fraudulent ID & Deepfake Benchmark Report 2026” report the same month, which draws on deepidv’s analysis of four million synthetic identities and more than 10,000 deepfakes.

Biometric injection attacks are up ninefold year-over-year, according to the report, which focuses on the scale and rapid growth of the AI fraud problem.

In May deepidv joined the Age Verification Providers Association (AVPA), which lists the company as a provider of age verification based on government-issued ID checks, reusable digital ID and facial age estimation.

Article Topics

biometrics  |  deepfake detection  |  deepfakes  |  deepidv  |  research and development  |  Scam.ai