The Alignment Problem, by Brian Christian

The Alignment Problem (2020)

The original formulation of the alignment problem and the connection between technical and moral questions. Since it was written before the LLM revolution, this book focuses on more conventional AI systems, which makes the conceptual structure easier to see. My book borrows the frame and points it at institutions rather than models.

Read it asking: One frequent critique of the x-risk school is that it focuses on theoretical or long term risks over real societal harms. Do you think there is a (historical) analogy to climate change?