Skip to content

Three programs at Ω Labs this fall: Frontier AI Governance (apply by Nov 10), AGI Strategy (apply by Nov 16) and AI Safety Technical Fundamentals (rolling enrolment) →

Three programs at Ω Labs · See the programs →

AI Safety Montréal
Join EN FR

AGI Strategy course

Five Monday evenings to get a clear view of where AI is heading, what could go wrong, what is being proposed to make it go well, and where you could help. Based on the AGI Strategy course developed by BlueDot Impact; we follow their curriculum and meet in person in Montréal.

Apply by November 16 →

Not sure it’s for you? Apply anyway.

When
Mondays, 6 to 9 PM, November 23 to December 21, 2026
Where
Ω Labs, 3813 rue Saint-Denis, Montréal
Time commitment
About 5 hours a week, in person every week
Cohort
10 to 12 people
Language
Materials in English, sessions in English or French depending on demand
Facilitator
Étienne Langlois
Cost
Free
Apply by
November 16 at 11:59 PM EST

Who it’s for

  • Professionals from any field, technical or not, who wonder whether their experience could be useful in AI safety.
  • Students and early-career people considering research, policy or engineering work on AI risk, who want to understand the whole field before picking a lane.
  • Anyone who has followed the AI risk debate from a distance and now wants a structured way in, with a group that meets in person.

It is an introduction: no previous course or technical background is needed.

The five sessions

Each week: do the unit’s readings and exercises before Monday, then three hours of discussion at Ω Labs.

Nov 23 Unit 1: The race to AGI

What is at stake as companies race to build AGI, what future you would want from it, and why competition makes coordination so hard. A 2026 case study: the Hugging Face incident.

Core readings include:

  • Daniel Selsam, “Personal Statement on AI Risk” (2026)
  • Holden Karnofsky, “Racing through a minefield: the AI deployment problem” (2022)
  • Peter Wildeford, “The Scramble: getting in position to pace the frontier” (2026)
  • Dwarkesh Patel, “The Rise and Fall of Agent Civilizations” (2026)

Nov 30 Unit 2: Drivers of AI progress

What has driven AI progress so far, what today’s systems can do on the evidence of benchmarks and trend data, and how progress could speed up once AI helps build AI.

Core readings include:

  • Ben Todd, “The case for AGI by 2030” (2025)
  • David Owen, “What will AI look like in 2030?” (2025)
  • Epoch AI, “Trends in Artificial Intelligence”
  • METR, “Frontier Risk Report (February to March 2026)” (2026)
  • Kokotajlo, Alexander, Larsen, Lifland and Dean, “AI 2027” (2025)
  • Cunningham et al., “The Economics of Recursive Self-Improvement” (2026)

Dec 7 Unit 3: Pathways to harm

Three ways the transition to advanced AI could go badly: losing control of AI systems, power concentrating in a few hands, and biological misuse. Then you weigh the arguments and form your own view.

Core readings include:

  • Holden Karnofsky, “AI Could Defeat All Of Us Combined” (2022)
  • Paul Christiano, “What failure looks like” (2019)
  • Ajeya Cotra, “Why alignment could be hard with modern deep learning” (2021)
  • Stix, Hallensleben, Ortega and Pistillo (Apollo Research), “The Loss of Control Playbook: Degrees, Dynamics, and Preparedness” (2025)
  • Hadshar, “Extreme Power Concentration” (2025)
  • Kevin Esvelt, “On Civilization-Threatening Bioweapons” (2026)
  • Guerra et al., “Building a Defense-in-Depth Biosecurity Strategy for the AI Era” (2026)

Dec 14 Unit 4: Strategies for the AGI transition

The plans on the table for getting through the transition: the technical problem of making advanced systems safe, the AI 2040 “Plan A” and its alternatives, and the objections raised against them. You finish with your own assessment of the situation.

Core readings include:

  • Larsen, Dean, Halstead, Lifland, Greenblatt and Kokotajlo, “AI 2040: Plan A” (2026)
  • Rohin Shah et al., “An Approach to Technical AGI Safety and Security” (2025)
  • Evan Hubinger, “Alignment remains a hard, unsolved problem” (2025)
  • Anthropic, “Teaching Claude why” (2026)
  • Eli Lifland, “AI 2040: Comparing Possible Plans” (2026)
  • Tom Davidson, “Plan A’s problem with dry tinder” (2026)
  • Richard Ngo, “Selective Optimism: a critique of AI 2040” (2026)

Dec 21 Unit 5: Start contributing

Choose where to focus, go deeper in one area, and leave with a personal action plan, plus, if you like, a one-page summary of what you want to do next.

Core readings include:

  • Neel Nanda, “Become a person who Actually Does Things” (2022)
  • Coefficient Giving, “Project Tailwind: Initiatives”
  • Dewi Erwan, “Personal Action Plan Template” (2025)

Each unit heading links to BlueDot’s unit page, where most units also list optional readings; a free BlueDot account unlocks the full course content.

Selection

Applications close November 16. We read every application and answer everyone shortly after the deadline. We look for people who will do the readings, come every week and act on what they learn.

FAQ

Do I need a technical background?
No.
English or French?
The readings are in English. Each session runs in English or French, depending on what the group needs.
Is this a BlueDot course?
It is our own course, built on the curriculum BlueDot Impact publishes. It is not an official BlueDot cohort.
Can I miss a session?
Tell us ahead of time. The course works best when everyone comes all five weeks.
Does it cost anything?
No, it is free.

Curriculum: the AGI Strategy course by BlueDot Impact, used under BlueDot’s guidance for independent cohorts. This is AI Safety Montréal’s own cohort, not an official BlueDot one. Readings belong to their authors.