Re: “AI coming at public service like a ‘freight train at a tremendous speed’ and it could mean thousands of job losses, says Savoie,” (The Hill Times, July 21, p. 35).
Professor Donald Savoie is right about the freight train, but wrong about the danger. What really threatens Canadians is a federal service model still running on steam power.
There are 2.2 million immigration, visa, and citizenship files that remain in Immigration, Refugees, and Citizenship Canada’s queue—officially acknowledged on Canada.ca. Last year, the Taxpayers’ Ombudsperson logged 2,796 complaints, most about being unable to reach a Canada Revenue Agency agent. Passport lines grew so long that Ottawa now offers a “30 business days or it’s free” guarantee just to restore public trust. When Employment Insurance tested a machine‑learning triage in July 2023, it processed 40,000 claims in weeks, and saved $2.6-million. That is not a pink‑slip apocalypse; it is proof that algorithms can bulldoze backlogs while freeing skilled officers to handle complex cases.
Ottawa’s payroll exceeds $55-billion annually. If artificial intelligence were to shave even five per cent of the rote work, the savings could fund faster veterans’ benefits, wildfire response, or dental care. Instead of fearing job losses, the Treasury Board should set a public target: every high-volume program will have an AI co-pilot by 2026, or departments will explain to Canadians why they’re still waiting.
Dr. Matt Crowson
[Canadian] Clinician‑data scientist, Harvard Medical School
Boston, Mass.