Launching soon — get early access:
Back to Glossary
Ethics & Safety

Constitutional AI(CAI)

Definition

An Anthropic-developed approach to training AI systems to be helpful, harmless, and honest using a set of principles rather than extensive human feedback.

In-Depth Explanation

Constitutional AI uses a set of written principles (a "constitution") to guide AI behavior during training. The model critiques and revises its own outputs based on these principles, reducing the need for human labelers. This approach helps create more transparent and consistent AI alignment, as the rules are explicit rather than implicit in training data.

Real-World Example

Claude is trained using Constitutional AI principles that include being helpful while avoiding harm and deception.

5 views0 found helpful

Related Terms

AI Alignment

The challenge of ensuring AI systems behave in accordance with human intentions, values, and goals.

AI Safety

The field focused on preventing AI systems from causing unintended harm and ensuring they remain beneficial.

Reinforcement Learning from Human Feedback(RLHF)

A training technique that uses human preferences to fine-tune AI models to be more helpful, harmless, and honest.