MAIA AI Safety Fundamentals

Every semester, MAIA runs a semester-long introductory reading group on AI safety, covering topics like neural network interpretability,¹ learning from human feedback,² goal misgeneralization in reinforcement learning agents,³ and eliciting latent knowledge.⁴ The program meets weekly in small groups, with dinner provided and no additional work outside of meetings.

See here for more details on the curriculum.

Apply here by Friday, Feb 17.

  • No experience with machine learning is expected, though we may group cohorts according to previous experience; we’d love to see people from a large variety of backgrounds with an interest in learning about AI safety.

  • AI Safety Fundamentals presents AI safety from a technical perspective, often reading research papers discussing specific parts of the alignment problem. We don’t discuss the governance implications of these problems, though expect this background to be highly useful context for later work on AI policy & governance.

    For the first time this semester, we are also running the Governance, Policy, & Strategy (GPS) Fellowship, which we expect to be even more useful to governance-oriented people.

  • Last semester, we received over a hundred applicants and accepted around half. We expect this semester to be similar.

  • We ask for your availability in the application, and will attempt to accommodate people’s schedules when forming cohorts.

  • We expect participants to go through the Week 0 content, which gives a basic introduction to machine learning.

  • If you’ve already read all the material in the curriculum, please email us at to discuss other ways of getting involved with MAIA.

Joint MAIA and HAIST retreats, where members and intro fellows discussed AI alignment and interacted with researchers from Redwood Research, OpenAI, Anthropic, and more.