Amii (Alberta Machine Intelligence Institute): The Complete Guide to the World Capital of Reinforcement Learning

Amii isn't just one of Canada's three national AI institutes—it's the world capital of reinforcement learning. With the 2024 Turing Award winner as Chief Scientific Advisor and DeepMind's only North American research lab, Edmonton has become where AI learns to do things.

What Is Amii?

Amii (Alberta Machine Intelligence Institute) is Canada's reinforcement learning powerhouse—an independent nonprofit that bridges world-class AI research with Prairie industry transformation. Founded in 2002 at the University of Alberta, Amii predates both Vector Institute (2017) and Mila's formal establishment (2017), making it the oldest of Canada's three Pan-Canadian AI Strategy institutes.

While Vector focuses on industry partnerships and Mila leads fundamental research, Amii dominates one specific domain: reinforcement learning—the branch of AI that teaches machines to learn from trial and error, like humans do.

Who Leads Amii?

Rich Sutton: The Father of Reinforcement Learning

Rich Sutton isn't just Amii's Chief Scientific Advisor—he literally invented reinforcement learning as we know it.

When Rich Sutton joined the University of Alberta in 2003, he brought reinforcement learning's center of gravity with him. Every graduate student studying RL worldwide has read his textbook. Every modern RL system—from game-playing AIs to robotics—builds on algorithms he pioneered.

"If we want real intelligence, AI needs to learn by doing, by trial and error. I think of myself as a reinforcement learning agent." — Rich Sutton, Chief Scientific Advisor, Amii

Sutton on AI Safety

Unlike many AI pioneers, Sutton takes a contrarian view on existential risk:

"The doomers are out of line and the concerns are overblown." — Rich Sutton, BetaKit interview, 2024

On large language models (LLMs) like ChatGPT and Claude:

"I don't think that's the direction that's going to lead to full intelligence." — Rich Sutton, Dwarkesh Podcast

His AGI timeline estimates? A 1-in-4 chance by 2030, 50/50 by 2040—faster than many expect, but through reinforcement learning rather than language models.

Michael Bowling: DeepMind Alberta

Michael Bowling bridges Amii and DeepMind, co-leading DeepMind Alberta while maintaining his University of Alberta professorship.

"AI technology isn't going to be a small thing; there's no way it can't change the world." — Michael Bowling, Canada CIFAR AI Chair, University of Alberta

Bowling's poker-playing AIs demonstrated something profound: AI could master games of imperfect information—where you can't see your opponent's cards—requiring reasoning about uncertainty, deception, and opponent modeling.

What Makes Amii Different from Vector and Mila?

Canada's three national AI institutes each carved out distinct identities:

Why Reinforcement Learning Matters

Reinforcement learning is fundamentally different from other AI approaches:

• Supervised learning (Vector's strength): Learn from labeled examples
• Unsupervised learning (Mila's strength): Find patterns in data
• Reinforcement learning (Amii's strength): Learn from trial and error

This matters because RL is how AI learns to act in the world—not just recognize patterns, but make decisions with consequences. Every robotics system, every game-playing AI, every recommendation algorithm that improves from user feedback—all built on Amii's foundational work.

How Does DeepMind Alberta Connect to Amii?

In 2017, Google DeepMind established its only North American research lab in Edmonton—not San Francisco, not Toronto, not Montreal. The reason? Rich Sutton and the University of Alberta's unmatched reinforcement learning talent.

DeepMind Alberta

DeepMind Alberta researchers maintain university affiliations, publish openly, and train the next generation of RL researchers—a model that keeps Edmonton's talent pipeline flowing while giving DeepMind access to world-leading expertise.

What Research Does Amii Lead?

Reinforcement Learning Foundations

Amii researchers have contributed core algorithms that power modern AI:

Recent Research Highlights

How Does Amii Support AI Startups?

Amii runs one of Canada's most active AI startup programs—completely free for founders.

Amii Startup Investment Program

Notable Amii Alumni

Upper Bound AI Conference

In 2025, Amii launched the Upper Bound AI Startup Pitch Competition—the first dedicated AI startup showcase for Western Canada.

How Is Amii Funded?

Amii operates on mixed public-private funding with significant recent increases:

2025 Funding Announcements

Pan-Canadian AI Strategy Allocation

Under Canada's $443M Pan-Canadian AI Strategy, Amii receives funding alongside Vector and Mila:

Amii's smaller allocation reflects Alberta's smaller tech ecosystem—but the impact per dollar is outsized given RL's foundational importance.

What Programs Does Amii Offer?

For Workforce Development

AI Workforce Readiness Initiative

Amii's most ambitious program aims to transform Prairie workforce skills:

AI Career Accelerator Program (AICAP)

For professionals pivoting to AI careers:

For Researchers

Amii Fellowship

Canada CIFAR AI Chairs

For Companies

Industry Partnership Program

How Can Researchers Join Amii?

Faculty Positions

Graduate Students

1. Apply to University of Alberta (CS or related)
2. Identify Amii-affiliated supervisor
3. Access Amii resources through your program

The University of Alberta's machine learning program—built around Sutton and his collaborators—is arguably the world's best for reinforcement learning.

Postdoctoral Positions

What Makes Edmonton Unique for AI?

The Edmonton AI Ecosystem

The Prairie AI Opportunity

Alberta's traditional industries—oil and gas, agriculture, logistics—face automation pressure. Amii bridges the gap:

The $3M PrairiesCan funding for 30 Prairie SMBs exemplifies this mission: taking world-class research and applying it to real business problems.

How Does Amii Approach AI Safety?

The Sutton Position

Unlike Yoshua Bengio's safety-first advocacy or Geoffrey Hinton's existential warnings, Rich Sutton represents the optimistic RL view:

• Current LLMs aren't the path to AGI — True intelligence requires learning from interaction
• Safety concerns are overblown — The "doomers" lack technical grounding
• Beneficial by design — RL systems learn what we reward them for

This creates productive intellectual tension within Canada's AI community—Sutton at Amii debates Bengio at Mila on timelines, approaches, and risks.

Safety Research at Amii

Despite Sutton's optimism, Amii does fund safety-relevant research:

What's Next for Amii?

2026 Priorities

Long-Term Vision

Amii's thesis: reinforcement learning is how AI learns to act in the world—and Edmonton is where that happens. As AI moves from pattern recognition to decision-making, Amii's foundational work becomes increasingly central.

"I don't think that's the direction that's going to lead to full intelligence." — Rich Sutton on LLMs, pointing toward RL as the path forward

How Can You Connect with Amii?

For Researchers

• Apply for Amii Fellowship or CIFAR Chair
• Attend Upper Bound conference
• Propose collaborative research

For Startups

• Apply to the accelerator (free, $250K in cloud credits)
• Access Amii's investor network
• Recruit from talent pipeline

For Companies

• Explore partnership tiers
• Send employees through AICAP
• Sponsor applied research projects

For Everyone

• Join AGI House Canada to connect with Amii researchers
• Subscribe to our newsletter for Edmonton AI ecosystem updates
• Attend our events featuring Amii faculty and alumni

---

Related Reading

Amii's Key People

• Rich Sutton: Father of Reinforcement Learning — Amii's star researcher and RL pioneer

Western Canada AI

• Vancouver: Canada's Gateway to Global AI Markets — Deep tech and the Pacific AI corridor

Canada's Other AI Institutes

• Vector Institute Complete Guide — Toronto's AI research and commercialization center
• Mila Complete Guide — Montreal's world-leading AI research center
• The AI Triangle: Canada's Three Superpowers — How Canada's AI hubs work together

---

Join AGI House Canada to connect with Canada's reinforcement learning community. Scan the QR code to join our WhatsApp group or subscribe to our newsletter.

Data current as of January 2026. Statistics from Amii public reports, CIFAR announcements, and government press releases.