A carrier moves faster than the regulator
On July 28, 2026, Bell announced it had become the first Canadian carrier to deploy AI-powered detection that identifies spoofed calls in real time — not just spam, but calls where the caller ID has been deliberately forged to impersonate a bank, a courier, or a government office. The feature rolls out automatically to Bell, Virgin Plus, Lucky Mobile, and affiliated brand customers on iOS and Android, no opt-in required. Since Bell's original Suspicious Call Detection launched in 2025, the network has analyzed more than 4.4 billion calls and blocked or labelled over 540 million as suspicious or fraudulent, according to Bell's press release. The company estimates the new model will catch 60-70% more suspicious calls than the prior system.
The target is specific: the CRTC attributes roughly 40% of the unwanted-call complaints it receives to caller ID spoofing and impersonation (MobileSyrup, citing CRTC figures). That single tactic — making a screen read "CRA" or "RCMP" — is the hinge on which Canada's version of the "digital arrest" scam turns: a spoofed call or a texted link opens into a video call with someone posing as a police officer, complete with a forged warrant on screen, who tells the target they face immediate arrest unless they pay by gift card or crypto. The Ontario Provincial Police say some victims have lost their life savings; RCMP detachments in Manitoba, Alberta, and BC have all issued separate warnings on the same pattern in the past year (per CBC). It is the same coercive-video-call template that has metastasized globally — including through WhatsApp — and that India's Supreme Court is currently pressing the RBI to address with a formal SOP for banks. Canada's version runs on the same psychological lever: a spoofed number lends the first call enough legitimacy that the victim answers, and enough authority that they stay on the video call instead of hanging up.
The case for making this mandatory, not optional
There's a real argument that carrier-level spoof detection shouldn't depend on which company a customer happens to subscribe to. Fraud is a negative externality — a Rogers or Telus customer who gets scammed is harmed by a gap in a competitor's network security, not just their own provider's. Seniors, recent immigrants unfamiliar with Canadian institutions, and non-native English speakers are disproportionately targeted by impersonation scams precisely because authority-mimicking caller ID is most effective against people least equipped to question it. A regulator could reasonably conclude that a protection this consequential shouldn't roll out first to Bell's ecosystem while other subscribers wait years for their own carrier to catch up — and that argument only gets stronger given how slowly the CRTC's own prior intervention has worked.
That prior intervention is Telecom Decision CRTC 2021-123, which made STIR/SHAKEN caller-authentication mandatory for all telecom providers as a condition of service, effective November 30, 2021. Nearly five years later, the CRTC's own guidance still concedes that STIR/SHAKEN protection is incomplete — it verifies only IP-based calls, and adoption gaps persist where carriers route traffic through legacy infrastructure. A mandate that took years to draft and still hasn't closed the gap is a weak advertisement for regulation-first sequencing.
Why the market got there faster
Bell's tool didn't wait for a CRTC order. It shipped because detecting fraud in real time, at the scale of billions of calls, is now something a carrier can do competitively — Bell is marketing "first in Canada" as a customer-retention feature, which is exactly the incentive structure regulators should want operating in the background. AI-based anomaly detection improves continuously as it ingests more traffic; a static technical mandate like STIR/SHAKEN, specified once and left to a governance authority, does not. The Canadian Anti-Fraud Centre logged over $704 million in reported losses across 112,000-plus reports in 2025 alone — a figure the agency itself says likely represents only 5-10% of actual fraud, meaning true losses may run into the billions (Canadian Anti-Fraud Centre). Against a threat moving that fast, a carrier that can retrain its detection model monthly has a structural advantage over a regulator that revises telecom orders on a multi-year cycle.
None of this means the CRTC should stand down. Its useful role now is ensuring parity — requiring competing carriers to hit comparable detection outcomes and publish results, the way it already requires STIR/SHAKEN readiness reporting — rather than dictating the specific model or vendor. Mandating outcomes while leaving the engineering to carriers racing each other for customers is the sequencing that got Bell's tool shipped in months instead of years. A prescriptive mandate for this exact AI approach, arriving in 2028 after consultation, would protect fewer Canadians than letting Bell's competitors feel obligated to match it now.