Resources

Understanding the risks

• AI 2027 (opens another website)

  A detailed, month-by-month scenario of how the arrival of superhuman AI could unfold, and where it could go wrong.

• 80,000 Hours: risks from advanced AI (opens another website)

  Why 80,000 Hours ranks reducing risks from AI among the world's most pressing problems, and how people can help.

Careers

• Free 80,000 Hours career guide (opens another website)

  Sign up to the 80,000 Hours newsletter and they'll send you a free book. The career guide is one of the options.

• 80,000 Hours career guide, online (opens another website)

  The full guide to finding a fulfilling, high-impact career, free to read.

Curricula

• BlueDot Impact: AGI Strategy (opens another website)

  A free 25-hour course on where AI is heading, how it could go wrong, and what can be done.

• ARENA curriculum (opens another website)

  Hands-on coding exercises for technical AI safety: PyTorch fundamentals, transformer interpretability, reinforcement learning and model evaluations.

Other reading groups in Montréal

• AI Safety Reading Group at McGill

  A reading group at McGill hosted by Prof. Adam Oberman and Dr. Mohamed Amine Merzouk, focused on mathematical approaches to AI safety. More details available via email.

• AI Safety Reading Group at Mila (opens another website)

  A research seminar at Mila where authors present their own AI safety papers, in person or online.