# Constitutional AI

> Source: https://sukruyusufkaya.com/en/glossary/constitutional-ai
> Updated: 2026-05-13T21:00:56.018Z
> Type: glossary
> Category: uretken-yapay-zeka-ve-llm
**TLDR:** An alignment approach that tries to guide model behavior through explicit principle sets and normative rules.

<p>Constitutional AI aims to guide model behavior not only through preference examples but through more explicit principled frameworks. This can make high-level goals such as safety, helpfulness, and harm reduction more systematic. It brings rule-based alignment thinking into large-scale model behavior design.</p>