Are We Racing Into AI With 20th‑Century Plumbing?
|
Executive Summary
|
|
Is Canada be expanding AI use faster than it is building the infrastructure, governance and skills to support it? Ottawa’s new “AI for All” strategy and National Council on Artificial Intelligence aim to boost adoption, domestic compute and safety, while Alberta courts “digital refinery” data centres and a $72B Emera–Canadian Utilities merger chases power demand.
If you haven't yet, check out our Atlas Dossiers for deep dives on different issues about Canada and AI.
Happy readings and a happy weekend!
|
|
|
|
Canada leads in AI talent but lags in computing power and access. Without better infrastructure and broad access, innovation and jobs may move abroad. |
|
|
|
AI use in organizations has surged, but most lack policies, training and clear strategies. CPA Canada urges stronger federal rules. Canadians should ask employers about AI guidelines and training to protect data, jobs and decision quality.
|
|
|
|
Canada’s AI health gains risk stalling without better digital infrastructure; HealthCareCAN urges funding for data systems, staff training, and secure tech so hospitals can safely adopt tools that improve patient care.
|
|
|
|
Alberta is courting big data centres with no subsidies, requiring firms to fund their own power links and follow strict siting rules, aiming to grow jobs while protecting grids, prices, and nearby communities—offering a model for future Canadian projects.
|
|
|
| The conversation around AI infrastructure and its economic impact often generates more heat than light. Nicolle Weeks breaks down the reality of data center regulations, the nuance of job disruption, and what "sovereign compute" actually means for domestic policy. |
|
|
|
Canada’s privacy commissioners found OpenAI’s ChatGPT 3.5/4 breached privacy rules; Bill C-36’s “legitimate interest” could weaken protections. |
|
|
|
Emera and Canadian Utilities plan a $72B merger to fund power for growing data centres, which could affect electricity demand, grid investment, and future power costs for Canadian homes and businesses.
|
|
🎙️ The Big Story
| October 08, 2026
Canada's billions in AI investments since Mark Carney's premiership rely on existing criminal codes for harms like the Tumbler Ridge shooting and Library and Archives hacking attempts, leaving developer accountability undefined amid rapid deployment.
There is a space here if Canada wants to ensure that safety is a core component of how these AI tools are developed in the country and deployed.
|
|
🎙️ Canada’s Entrepreneur
| October 08, 2026
Canadian small-business AI tools gain traction by solving missed opportunities like off-hour calls for realtors and restaurants rather than pushing technology adoption itself, with founders iterating through direct owner feedback in the GTHA before national scaling.
|
|
🎙️ The Flourish Feed Podcast
| October 08, 2026
Canadian organizations risk encoding employee expertise into digital twins without resolving ownership, liability or accountability, especially in large legacy sectors where intuition and scenario-based judgment exceed standard operating procedures.
"Canada is ranked in AI training and adoption, so more adoption. So there's a gap, but this is a gap where women can leap on"
|
|
"Je suis honoré d’avoir été sélectionné par le premier ministre @MarkJCarney pour rejoindre le nouveau Conseil national sur l’intelligence artificielle, qui aura pour mission de soutenir le Gouvernement du Canada afin de rendre l’IA plus sécuritaire, protéger notre démocratie et façonner une trajectoire de l’IA plus bénéfique pour la population.
[https://t.co/5dMKgb57Bq]
***
Honoured to have been selected by Prime Minister @MarkJCarney to join the new National Council on Artificial Intelligence, which will support the Government of Canada on making AI safer, protecting our democracy, and shaping a more beneficial AI trajectory for Canadians.
[https://t.co/ueJqdfs9YU]"
@Yoshua_Bengio
💡 Canada forms a national AI council to guide safe, democratic, citizen-focused AI policy
|
|
"BREAKING: PM Carney just launched the National Council on Artificial Intelligence.
Independent advice on AI for All — adoption, national champions, sovereign compute, safety, and democracy.
Members include Yoshua Bengio, Michelle Zatlyn, Ajay Agrawal, Sanja Fidler and more.
Prosperity and sovereignty in the age of AI belong to nations that build and govern it on their own terms. That's the right frame for Canada.
https://t.co/VoVNAy1plW
#ai #canada #cdnpoli"
@iambasitshah
💡 Canada is formalizing sovereign AI strategy with expert-led national governance and adoption
|
|
"Mecka AI, a Toronto- and New York City-based physical AI startup, has raised a $60 million USD Series B led by Sequoia Capital, with backing from Nvidia, Microsoft, Qualcomm, Samsung, Kindred Ventures, Framework Ventures, Neo and other investors. The company is building the data and deployment layer for robotics, collecting real-world sensor and first-person video data to train humanoid and other robots, and plans to use the new capital to scale its data infrastructure, expand its research lab and grow commercial robot deployments .."
@TradedVC
💡 Mecka AI raises $60M to build data and deployment stack for humanoid robotics |
|
".@Mila_Quebec's Alex Shee shares how Canada can keep its #AI talent https://t.co/9ta7mu6rMl"
@BetaKit
💡 Canada needs stronger AI retention strategies to keep top talent from leaving
|
|
"Building autonomous vehicles starts with cybersecurity.
We worked with @UWaterloo & @Transports_gc to test #AI systems against cyber threats to help keep future transportation safe & secure.
https://t.co/d0jljObvt9
#NRCAnnualReport #CyberMonth2026 #NRCDigitalTech https://t.co/hI2ujB1deU"
@NRC_CNRC
💡 Secure-by-design AI is essential for safe autonomous and future transportation systems
|
Selected AI Research from Canada
|
|
|
|
Mila - Quebec Artificial Intelligence Institute
| October 06, 2026
|
|
A single malicious image can “poison” an AI assistant’s memory so its influence lingers even after the image is removed. Later, a specific trigger question can make the model follow hidden instructions. This shows AI chatbots can be secretly reprogrammed via long-term internal memory, not just visible prompts.
|
|
|
|
Mila - Quebec Artificial Intelligence Institute
| October 07, 2026
|
|
RoboJEPA is a giant “imagination engine” for robots: trained on data from 12 different robot types, it predicts future situations and plans actions. Its accuracy improves in a predictable way as you add more computing power, and it can control real robots directly from just a single goal image. |
|
|
|
University of Regina
| October 05, 2026
|
|
Shows how popular open AI models respond to Indigenous mental health scenarios. Finds they almost always mention crisis hotlines but rarely say “I’m not a doctor” or suggest seeing a clinician, and their language is harder to read than recommended, raising real concerns about safety, clarity, and cultural fit.
|
|
🤔 Question of the Week
Where do you think Canada's AI policy should lean into at this moment? More Speed or safety?
|
|
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
October 09, 2026
202-2149 Yonge St. Toronto ON, M4S 0C5
|
|