Home / 🛡️ Safety / Constitutional AI
🛡️ Safety active v1.0

Constitutional AI

by Anthropic
Self-supervised alignment method where AI systems evaluate and revise outputs against a set of principles (a constitution).

Key Features

  • alignment
  • self-supervision
  • principles

📄 Specification

Category 🛡️ Safety
Version1.0
MaintainerAnthropic
Status active
Sourcearxiv.org/abs/2212.08073

💻 Code Examples

Coming soon — integration guides and code examples for Constitutional AI.
Want to contribute? Submit an example →
Discovered March 24, 2026
Last updated March 24, 2026