A proposed open standard for declaring AI content usage preferences (ai.txt, meta tags, headers)
AIP-1 (Artificial Intelligence Preference Protocol) is a proposed open standard that allows websites to declare how their content may be used by AI systems.
It acts as a natural extension of existing web conventions—bringing clarity and transparency to AI-era content usage.
AIP-1 is designed to:
- Provide a simple, universal way to express AI usage preferences
- Enable transparent interaction between publishers and AI systems
- Support responsible model development without restricting innovation
- Create a common language across the web ecosystem
The protocol introduces three lightweight components:
A root-level file (similar to robots.txt) defining site-wide preferences.
Fine-grained, page-level control over AI usage.
Support for APIs and dynamic content delivery.
Sites and services that adopt AIP-1 can display a public badge to signal compliance.
[](https://github.com/SquiddoTech-Systems/AIP-1/)<a href="https://github.com/SquiddoTech-Systems/AIP-1/"><img src="https://img.shields.io/badge/AIP--1-Compliant-brightgreen" alt="AIP-1 Compliant Badge" style="border: none; display: block; max-width: 100%; height: auto;"></a>As AI systems become a core part of the web, clear communication between:
- Content creators
- Platform operators
- AI developers
becomes essential.
AIP-1 aims to provide that missing layer.
This project is intentionally:
- Neutral (not anti-AI)
- Compatible with existing standards
- Practical for real-world adoption
It is designed with potential adoption in mind by organizations such as OpenAI and Google, as well as the broader web community.
Draft Proposal — Early Stage
This project is open for feedback, iteration, and collaboration.
- Publish full specification (AIP-1 spec)
- Gather community feedback
- Explore standardization pathways
- Build reference implementations
Contributions, ideas, and discussion are welcome.
MIT License