Canada Bets a lot on AI.
|
Executive Summary
|
|
AI infrastructure and health care dominated Canadian AI developments this week: B.C. will allocate 400 megawatts over two years to new AI data centres amid public protests, while a New Brunswick MP urges the province to reject a proposed AI data centre unless it fully covers its costs. Radical Ventures closed a $1‑billion Breakouts Fund backed by pension giants to scale Canadian AI firms.
In health care, Ontario providers are being warned to address privacy, safety, human rights and governance before deploying AI, Montreal experts debated whether AI can ease Canada’s doctor shortage, and AlayaCare outlined an agentic AI roadmap for home-based care. Canada’s AI research and industry ties were also in focus at the ALL IN 2026 event in Montreal.
|
|
|
|
B.C. will reserve 400 MW of power over two years for new data centres, with winning bids announced this month. This could affect local power use, bills, and land. Watch for public consultations and project details in your area.
|
|
|
|
Ontario health providers are urged to assess privacy, patient safety, bias, consent and oversight before using AI tools, and to set clear policies, human review and accountability to meet legal duties and protect patients.
|
|
|
|
Radical Ventures raised a record $1B Breakouts Fund to back Canadian AI scaleups, with major pension funds investing. This could mean more local tech jobs and investment opportunities tied to homegrown companies like World Labs and You.com.
|
|
|
|
A New Brunswick MP urges the premier to reject a proposed Beacon AI data centre over power use and cost concerns. The province says any project must fully cover its grid costs so electricity rates for residents and businesses do not rise.
|
|
|
|
Canada’s ALL IN 2026 will spotlight national AI strengths, deepen ties with Germany and global partners, and showcase Mila and firms like Hypertec—creating more research, jobs and tech investment opportunities for Canadians.
|
|
|
|
Experts at a Montreal AI conference debated using symptom-recognition tools to ease Canada’s doctor shortage, exploring ways tech could speed diagnosis, reduce wait times, and support overburdened clinics while stressing safety and oversight.
|
|
|
|
AlayaCare revealed a new AI roadmap at its Toronto conference, aiming to help home-care providers automate routine tasks and improve patient support, which could mean faster, more reliable in‑home care services for Canadians.
|
|
🎙️ Disruptors
| September 16, 2026
Canada pioneered deep learning via sustained public investment yet now faces selective US access to frontier tools, creating cyber risks and uneven capabilities. Leaders advocate sovereign models like Cohere and democratic supply chains to retain control over deployment and capture returns from domestic innovation.
We invented it. Our taxpayers invested in it. Our taxpayers bet on deep learning before when no one else believed in it.
|
|
🎙️ Connect Newsroom
| September 15, 2026
Canadian pension boards including CPPIB are allocating to AI and quantum at an investment summit where speakers stated rules for the most consequential technology cannot be set by commercially aligned companies alone.
If this is the most consequential technology in human history, then the rules for it cannot be written by a small group of commercially aligned companies.
|
|
🎙️ Beneath the Law
| September 15, 2026
A Canadian lawyer received a six-month suspension after submitting AI-generated fictitious case law in a family proceeding, then repeatedly misrepresenting the source to the court and Law Society during investigation.
|
|
"As AI gets more powerful, safety cannot be an afterthought. Today, Canada and Germany announced a combined $300 million investment in @LawZero_ , founded by @Yoshua_Bengio.
Canada helped lead the world in AI research. Now, we’re helping lead the work to make AI safe, reliable and worthy of trust.
🇨🇦
.."
@EvanLSolomon
💡 More governments are funding AI safety as a core priority, not an optional add-on
|
|
"Saskatchewan just quietly landed one of the biggest private-sector investments in Canadian history.
The province and Bell Canada are expanding the Regina-area AI data centre from 300 megawatts to a path toward 1.2 gigawatts, with more than $50 billion in associated capital investment. Bell AI Fabric is putting its national headquarters in the Regina area.
This works for Saskatchewan because the province actually has what the AI boom needs: land, reliable power, a cold climate, and a government that demanded Canadian ownership, local jobs, data sovereignty, and “bring your own power.” The extra capacity is funded by Bell, not by ratepayers. Construction jobs, permanent operations jobs, a head office, and local contractors are already part of the deal.
It works for Canada because compute is now strategic infrastructure. A Canadian-owned 1.2 GW hub, run by a Canadian company under Canadian rules, means we stop renting our digital future from someone else. When the data stays here, the rules stay here.
The Prairies used to be the place people flew over. Turns out space, power, and winter were the features, not the bugs. Wheat fields still matter. Now watt fields do too."
@legallymsjuno
💡 Canada is building sovereign AI compute at scale, with Regina as a 1.2 GW strategic hub
|
|
"OPENAI IS EXPLORING DATA CENTER PARTNERSHIPS IN CANADA, CITING THE COUNTRY'S ABUNDANT ENERGY, LAND AND GROWING AI INFRASTRUCTURE OPPORTUNITIES UNDER PRIME MINISTER MARK CARNEY'S STRATEGY."
@FirstSquawk
💡 OpenAI eyes Canada data centers, drawn by energy, land and pro-AI national strategy |
|
"ICYMI: Canada's #AI strategy is focused on growth, investment and creating 250,000 jobs by 2031, but what happens to workers whose jobs disappear?
Read: https://t.co/tdT0EcIybH"
@CIGIonline
💡 AI growth plans must address job loss and worker transition, not just investment and targets
|
|
"Canadians can sign up for the government's free AI basics course 💻
https://t.co/xJNYTKhDeR"
@DailyHiveVan
💡 Canada is upskilling citizens in AI basics, signaling broader national AI readiness efforts
|
Selected AI Research from Canada
|
|
|
|
McGill University
| September 13, 2026
|
|
Shows how AI systems could talk people into bad decisions in high‑stakes AI labs—like approving unsafe code, weakening safety checks, or bypassing oversight. Maps out five concrete “persuasion attack” scenarios and offers a step‑by‑step playbook for testing, measuring, and reducing these manipulation risks.
|
|
|
|
University of Calgary
| September 14, 2026
|
|
AI tools don’t just force programmers to learn new software; they also create stress about changing job roles, accountability for AI-generated code, faster work expectations, and long‑term career security, affecting how developers work today and how they plan their futures.
|
|
|
|
University of Calgary
| September 17, 2026
|
|
AI tools are helping doctors spot and treat strokes faster, but they can be opaque and unfair, especially across races and income levels. The paper shows current systems rarely test explanations or bias, and calls for unified standards so stroke AI is transparent, fair, and reliable for all patients.
|
|
|
|
University of Toronto
| September 16, 2026
|
|
Teens using AI tools almost never double-check answers, risking blind trust. Researchers built a way to measure how carefully students question, test, and cross-check AI outputs, identified key habits like “deep logical auditing,” and found four user types—from naïve trusters to strategic skeptics—to guide safer, smarter AI-based learning.
|
|
|
|
Toronto Metropolitan University
| September 12, 2026
|
|
AI in food delivery apps feels helpful but “stuck” to Gen Z: it repeats old choices instead of adapting to changing moods, health goals, or identity. Young users trust it only when it’s transparent, respects their autonomy, supports future goals, and clearly isn’t just pushing sales.
|
|
|
|
University of Saskatchewan
| September 13, 2026
|
|
When people co-write documents with AI help, they prefer seeing more detail about each other’s AI use—when it was used, what it generated, and how prompts were written. Sharing this information builds trust and coordination, so future writing tools should make AI involvement clearly visible to all collaborators.
|
|
|
|
Athabasca University
| September 12, 2026
|
|
Shows managers shouldn’t simply “trust” or “distrust” AI. Instead, they must match how much they rely on AI to how reliable it is in that specific decision, considering their own expertise, how easily results can be checked, uncertainty, and the real-world stakes involved.
|
|
🤔 Question of the Week
As AI data centres and healthcare tools scale across Canada, what governance model best balances innovation, energy use, and public trust?
|
|
Grow the Community
You are receiving this email because you have subscribed to our newsletter. Please help grow Canada's AI community to 10,000+ professionals!
Share link: www.northernsignal.ca
|
|
|
NorthernSignal.CA - Your Weekly Update
September 18, 2026
202-2149 Yonge St. Toronto ON, M4S 0C5
|
|