Constitutional AI is Anthropic’s method for training a model to follow a written set of principles by having it critique and revise its own answers, instead of relying mainly on human-labeled examples. It is one of the techniques used to make Claude safer and more predictable. This hub explains how the approach works and where it fits among AI safety methods.
Constitutional AI is Anthropic's method for training Claude to follow a written set of principles by critiquing its own answers, instead of relying mainly on human-labeled examples.
Read more →A paper honored at ICML 2026 argues that AI alignment techniques such as RLHF and Constitutional AI can double as tools for state censorship and political control.
Read more →