What is prosaic alignment?

Prosaic AI alignment is an approach to alignment research that assumes that future artificial general intelligence (AGI) will be developed "prosaically" — i.e., without "reveal[ing] any fundamentally new ideas about the nature of intelligence or turn[ing] up any 'unknown unknowns.'" In other words, it assumes the AI techniques we're already using are sufficient to produce AGI if scaled far enough. Because of this assumption that future AI systems will resemble current ones, prosaic alignment research is often relatively concrete and empirical, evaluating the effectiveness of proposed alignment techniques by trying them out on existing AI systems.

Some examples of prosaic alignment proposals are debate, imitating humans, preference learning, and iterated distillation and amplification.