Alberta's AI code audit just killed the best excuse Northwestern Ontario institutions had for not looking

The province scanned 466 million lines of legacy government code in about 20 hours using Claude, then gave the method away for free. That doesn't make modernization free for a Thunder Bay institution — but it removes the reason to never even start.

Alberta's government just did something worth Northwestern Ontario's attention for reasons that have nothing to do with Alberta's budget. Its Ministry of Technology and Innovation pointed roughly 50 Claude Code agents at its own aging systems and scanned 466 million lines of government code in about 20 hours — work the team estimates a traditional review would have taken around 6.5 years to complete. Then it published the entire method as 21 free, open-source white papers so any other government can copy it. The headline is not that Alberta has resources the Northwest doesn't. It's that the actual bottleneck — the years of specialist labour it used to take just to find out what shape your systems are in — has been shown, at real government scale, to collapse to a fraction of that. For a Thunder Bay municipality, health authority, school board, or college running the same kind of unreviewed legacy code, "we don't have the budget to even look" just got a lot harder to justify.

What Alberta actually did

The scan covered systems across all 27 of Alberta's provincial ministries; the Ministry of Technology and Innovation says it runs more than 1,280 applications and 3,400 collections of code in total. About 50 Claude Code agents worked autonomously and in parallel — first flagging known vulnerability patterns with a rules engine, then reviewing each flag and citing the exact file and line so a human developer could verify it. Where the scan found real problems, Claude Code often went further: generating fixes, writing tests that didn't exist, and in some cases rebuilding outdated systems in more current languages. Alberta's own estimate for modernizing that technology the traditional way — not just the initial scan, the full fix — is roughly $2 billion and more than a century of work.

Then it gave the method away

This is the part that matters most for a smaller jurisdiction. Alberta didn't just do the work quietly and keep the advantage. It published 21 technical white papers — the Velocity White Papers, at thevelocitywhitepapers.com — free, open-source, with step-by-step instructions, specifically so other governments facing the same aging-systems problem don't have to redo the R&D from scratch. Alberta's Minister of Technology and Innovation, Nate Glubish, put it plainly: "Albertans trust their government with some of the most sensitive information in their lives, and it is our responsibility to protect it. By using AI to find and fix vulnerabilities across our systems, we accomplished in hours what would have taken a traditional approach years to complete. This is what responsible government looks like in the AI era, and the best is still ahead of us." The province is also planning to use AI agents to consolidate 165 separate legacy systems into 16 new applications it will own outright, rather than continuing to license and patch the old ones.

Why this is a Northwestern Ontario story

Swap "Alberta ministry" for a City of Thunder Bay department, a district health authority, a school board, or a college, and the underlying problem is identical, only smaller in scale: systems built up over decades, patched by whoever was available at the time, with no complete picture of what is actually running or how exposed it is. That kind of review has always been theoretically possible and practically impossible — no NWO municipality or public institution has the budget to hire a team for a multi-year manual code audit. Alberta just demonstrated, at a scale no one can dismiss as a toy example, that the practical-impossibility half of that equation has moved substantially. The playbook for how it did that is now public and free to read.

A practical first step

The realistic move for an NWO institution is not to attempt Alberta's scale — it's to read what Alberta published before committing to a vendor or a system rebuild sight-unseen. The white papers are public at Alberta's own site, and using them to scope one system or one repository as a pilot is a cheaper, lower-risk starting point than commissioning a study from zero. From there, loop in whoever actually maintains your systems — an internal IT lead, if you have one, or a local implementation partner if you don't — and treat the pilot's findings, not a vendor's pitch, as the basis for what to fund next. If funding is the gap, confirm current eligibility directly with the relevant program before assuming it applies to a systems-review project; see thunderbayai.com's funding page for what is currently active in the Northwest.

Sources: Government of Alberta uses Claude to find and fix cybersecurity vulnerabilities across government systems — Anthropic, July 2026 (anthropic.com/news/alberta-government-claude-cybersecurity) | Alberta is open-sourcing its AI playbook for government — BetaKit, July 7, 2026 (betakit.com/alberta-is-open-sourcing-its-ai-playbook-for-government/) | Alberta government uses Claude to check its code — The Logic, July 2026 (thelogic.co/briefing/alberta-government-uses-claude-to-check-its-code/) | wire coverage syndicated across Alberta local outlets including Red Deer Advocate and Pipestone Flyer, July 6, 2026. The Velocity White Papers are published at thevelocitywhitepapers.com. Figures (466 million lines, ~20 hours, 6.5-year estimate, $2 billion/century modernization estimate, 165-to-16 system consolidation plan, the Glubish quote) are Alberta government and Anthropic's own reporting, cross-checked across the outlets above; no independent audit of these figures has been published. This is commentary, not a recommendation of any specific vendor, tool, or funding program — confirm any program's current eligibility directly with the program.