Marc Lijour

My new article on Canada's ill-fated Artificial Intelligence and Data Act: What was the AIDA, why did it fail to become law, and what's next for Canadian AI regulation amidst a turbulent political landscape? Canada's inability to pass comprehensive AI legislation is a unique and major failure of national AI policy. Researchers, practitioners, and policymakers in Canada and around the world should study this case closely. Thank you to Montreal AI Ethics Institute for inviting me to contribute this piece! https://lnkd.in/e2u5yZDm

114

6 Comments

"Canada remains stuck in the wrong AI paradigm. Ottawa’s latest AI investment package focused almost entirely on compute infrastructure, as if the biggest challenge was a lack of servers. Meanwhile, little to no funding was allocated to fostering AI adoption or literacy. That’s like building libraries while neglecting to teach Canadians how to read." My op-ed this morning in the Globe and Mail: https://lnkd.in/g3JW5mSH

135

18 Comments

Just had an article in the Financial Post. It’s a call to rethink Canada’s role in the AI race — not as a builder of the biggest model, but as the global guardian of AI safety. Remember when BlackBerry ruled the boardroom? Not because it had the flashiest features — but because its crypto was bulletproof. Canada has done this before. We can do it again. We missed the model gold rush. That’s fine. The next big boom? AI security. And we’re sitting on the talent, the trust, and the playbook. Instead of trying to out-ChatGPT, maybe we lead where it really counts: trust, resilience, and infrastructure Made-Secure-in-Canada. Check out the piece — and let me know: could Canada be the brakes the AI world actually needs? Special thanks to Canadian Cybersecurity Network and Francois Guay for amazing support for this article! #AI #Cybersecurity #CanadaFirst #TrustworthyAI 

72

7 Comments

Opennote (YC S25), a student-founded edtech startup operating between Guelph, Ontario and Irvine, California, has raised $850,000 USD (approximately $1.2 million CAD) in a pre-seed funding round. The round, structured entirely through SAFEs, was led by Afore Capital with participation from Sancus Ventures, Script Capital, untitled ventures, and angel investors. 𝗥𝗘𝗔𝗗 𝗧𝗛𝗘 𝗗𝗘𝗧𝗔𝗜𝗟𝗦 👉 https://lnkd.in/dNGX4ceh Ashik Mathew Tharakan Matthew Beigi Rishi Srihari Hasan Baig Vedant Vyas Rudra Chavda Abhigyan Arya Amer Aouameur Tahasin Shadat #Startups #Founders #FundingNews #VentureCapital #EdTech #PrivateEquity

102

3 Comments

Canada just opened a national consultation on its next AI strategy, running October 1–31, 2025. I’m sharing a two-part breakdown of my submission and a video series to match. Part 1 covers how Canada can turn clean power and cold climates into sovereign, low-cost compute and real productivity gains for industry, SMEs, and solo builders, with practical steps leaders can act on now. https://lnkd.in/eC6wd_iW Key points from Part 1 of my submission: • Retain a research edge by building provincial edge data centers next to hydro, knitting them into a sovereign fabric, and pricing inference far below U.S. retail so Canadian teams can iterate daily without burning cash. • Compete globally by exporting Canadian-run inference and industrial AI services, not just papers; treat compute like a public utility and NL as a pilot node. • End the campus–company split: embed AI researchers in firms for one to two years with reciprocal lab days; align IP so joint work ships, and default non-strategic components to open source. • Keep research competitive and ethical by pairing cheap, clean compute with strong privacy defaults, Canadian data residency for public workloads, and transparent model evaluations tied to deployment context. • Attract talent on life and mission, not salary: guaranteed low-cost compute, fast family immigration, portable provincial appointments, and the chance to build national-scale systems. • Focus adoption where Canada actually makes things: energy, fisheries, logistics, manufacturing, health systems, and northern services; put near-prem inference plus a visiting “AI resident” into clusters of 10–20 firms. • The main barrier is culture, not tooling: leaders must use AI daily before they mandate it; pair that with twelve-month inference credits and hands-on coaching for SMEs. • Measure what matters: domestic inference consumption per SME, export revenues from Canadian-run compute, time-to-pilot for embedded researcher projects, and cost-per-task declines in government and regulated industries. • Commercialization with sovereignty: stand up national and provincial compute for inference, price it 5–10× cheaper for Canadian companies, and offer Universal Basic Compute credits residents can use or sell. • Fix incentives: shift funding from slow degrees to rolling workplace bootcamps, micro-apprenticeships, and adoption grants; prioritize individuals and tiny teams alongside anchor firms. • Create more industrial champions by placing practitioner mentors in companies for one-hour and one-day interventions that unblock adoption, then scale what works across sector cohorts. • Learn from others without selling the steering wheel: welcome friendly capital while keeping control of critical compute, data residency, and safety standards on Canadian soil.

27

9 Comments

Graham 📊 Dobbs and my report on Canadian AI compute capacity was cited in Mark Carney’s “New Approach to Fiscal Management and Growth for Canada — We need a government that spends less, so Canada can invest more": "Canada's available compute resources—including its data centres, cloud, AI systems and supercomputers—are notably behind leaders such as the United States, Japan, and China, constraining the critical computing power essential for driving innovation. [...] To build more competitive manufacturing and services sectors, Canada must invest to become a clean energy superpower in nuclear, hydro power, wind, hydrogen, battery storage and carbon capture. To lead the AI revolution, we need to capitalise on our unique opportunity to build competitive data centres and intelligence infrastructure that are wired into the largest AI market in the world." Our report preceded the previously announced $2 billion federal AI investment and the Government of Canada’s proposed $15-billion to match $30 billion in pension fund investment in green data centres is far more important. Combined with Pierre Poilievre’s support for leveraging Canada’s energy sector to increase our data centres, Carney’s proposal captures a bi-partisan consensus that we need to invest more in public energy and data infrastructure. Canada’s greatest opportunity for overcoming our reliance on resources is building a sustainable and independent innovation sector with AI at its core. Our unique access to clean energy and land and history of AI leadership presents an opportunity to convert these resources into a data and tech advantage. Ensuring our data centres remain in the public interest will be crucial including local resource governance, social economic and labour market benefits; open and equitable access; and ethical data governance and stewardship. Thanks to The Dais at Toronto Metropolitan University for the collaboration across their ongoing AI research spanning: infrastructure, skills and jobs, governance, and adoption. Marium Hamid Viet Vu Mark Hazelden Thanks to those who’ve supported our work on AI compute and governance including Henri Vilandré Mila - Quebec Artificial Intelligence Institute; Marc Andrew IREN Daniel Brock Hon. David Lametti Alex Steinhouse Fasken Aftab Erfan Fergus Linley-Mota SFU's Morris J. Wosk Centre for Dialogue Brandon Jackson Joshua Tan Public AI Matthew Victor Elissa Strome Tibor Turi, PhD P.Eng. CIFAR Tony Douglas, Jr. OECD.AI Akbar Nurlybayev CentML Nish Sinnadurai Cerebras Systems Artem Mamzhiev Ben Davies Vector Institute Colin McKay Waabi Tagging Honourable Michelle Rempel Garner, P.C., M.P. Nathaniel Erskine-Smith for their awareness. See below my recent photograph of the Barcelona Supercomputing Center housed in a former church.

70

2 Comments

90% of Canadian public servants feel unprepared for AI adoption - here's the framework changing that. I recently came across interesting research about a new way to bring AI into our government services - it's called EPIC (Education, Partnership, Infrastructure, and Community). Working in AI projects with the Canadian Federal Government, I've learned that good planning makes all the difference. EPIC breaks down the complex world of AI into four simple parts that work together: • Teaching teams about AI capabilities • Building strong partnerships • Setting up reliable tech systems • Getting citizens involved The framework tackles real challenges we face daily: • Keeping data safe and private • Meeting security standards • Managing information properly • Understanding what AI can and can't do What I really like about EPIC is how practical it is. Instead of just theory, it gives clear steps for government teams to follow. It helps build trust while making sure AI actually helps people. This research is a game-changer for anyone working to bring AI into government services - whether you're a public servant, researcher, or industry partner. Share your experiences: How is your organization preparing for AI integration in public services? source: https://lnkd.in/ekXWQRM2

40

3 Comments

The AI economy won’t wait for outdated policy. At Upper Bound by Alberta Machine Intelligence Institute (Amii) Jess Sinclair from Council of Canadian Innovators | Conseil canadien des innovateurs (CCI) made it clear: We’re moving into a world where intangible assets and frontier firms define value — and the policy playbook needs a rewrite. Council of Canadian Innovators | Conseil canadien des innovateurs (CCI) is a national business council representing Canada’s fastest-growing tech companies. It advocates for smart public policy on talent, capital, data, and digital infrastructure to help domestic firms scale globally. CCI serves as a bridge between innovators and government to shape Canada’s innovation economy. It’s not about more or less government. It’s about smarter government.

53

2 Comments

📚 A Retail CBDC Design for Basic Payments: Feasibility Study A recent Bank of Canada feasibility study offers exciting perspectives on designing a retail central bank digital currency (#CBDC) optimized for basic payments. Here are the essential takeaways: 1️⃣ Micro-Partitioned UTXO Architecture: Identified as ideal for basic payments, offering #scalability, #speed, and #privacy. 2️⃣ Scalability: Systems like OpenCBDC demonstrated impressive scalability, achieving up to 250,000 transactions per second; more than adequate for national-scale #adoption. 3️⃣ Privacy-Enhanced Auditing: Privacy-preserving techniques could ensure transaction confidentiality without compromising #auditability. 4️⃣ Non-Repudiation Guarantees: Strong mechanisms were outlined to securely prove transaction integrity and settle disputes effectively. 5️⃣ Two-Tier Deployment Model: Clearly separates central bank functions (issuance, settlement) from intermediary roles (customer-facing activities), enhancing #efficiency and #compliance. 6️⃣ Further Exploration Needed: Areas such as system integration with existing retail #payment infrastructure and resilience to core system outages were highlighted for deeper investigation. As we shape the digital money landscape, findings like these will guide #policymakers and #innovators alike. 🔗 For more details, check out the entire paper here: https://lnkd.in/dPbxwP8r

25

OpenAI and other big tech companies are starting to roll out the next wave of artificial intelligence, designed to operate with more autonomy. Let’s down how agentic AI works and why some think it will change how you use the internet. Should we trust them with our data? Experts caution that with AI assistants, even tiny slip-ups can be costly. #agenticai https://lnkd.in/gnmfsvZr

27

3 Comments