The IETF standard for declaring AI usage preferences on web content
AIPREF (AI Preferences) is an emerging standard developed by the Internet Engineering Task Force (IETF) that enables content owners to declare preferences about how automated systems and AI models may use their content. The standard provides a standardized vocabulary and attachment mechanism for expressing these preferences through HTTP headers and robots.txt.
As AI systems increasingly consume web content for training and inference, content creators need a standardized way to express their preferences. AIPREF fills this gap by providing a simple, machine-readable format that AI systems can discover and respect.
The AIPREF standard is being developed by the IETF AIPREF Working Group, which brings together stakeholders from across the internet ecosystem including content creators, AI developers, web platform operators, and standards experts.
The AIPREF Working Group is chartered to develop standards for expressing preferences about automated processing of web content, with initial focus on AI training and generative AI systems. The working group aims to balance the interests of content creators, AI developers, and users while promoting transparency and control.
The AIPREF standard consists of two primary specifications, along with supporting RFCs:
Defines the four preference categories (bots, train-ai, train-genai, search) and their semantics. This specification establishes the vocabulary that content owners use to express their preferences.
Read Vocabulary Specification →Specifies how preferences attach to content through HTTP headers and robots.txt directives. This specification defines the Content-Usage header field and its integration with the robots exclusion protocol.
Read Attachment Specification →AIPREF provides content creators with a standardized, machine-readable way to express preferences about AI usage. This is particularly important for protecting creative works, proprietary content, and premium resources while maintaining control over how content is used.
AI developers benefit from clear, standardized signals about content usage preferences. AIPREF eliminates ambiguity and provides a consistent mechanism for discovering and respecting content owner preferences across the web.
A standardized preference mechanism promotes transparency, trust, and healthy relationships between content creators and AI systems. AIPREF helps ensure the web remains an open platform while respecting content owner rights.
AIPREF.dev is a free, open source tool for generating AIPREF-compliant configurations. This tool is designed to make implementing the AIPREF standard easy and accessible for everyone.
AIPREF.dev is an independent community tool and is not an official IETF website. This tool implements the AIPREF specifications developed by the IETF AIPREF Working Group. For official information about the AIPREF standard, please refer to the IETF resources linked above.
The AIPREF specifications are currently in draft status within the IETF standards process. The standard is under active development with input from stakeholders across the internet ecosystem.
As an emerging standard, AIPREF adoption is growing among content platforms and AI developers. Early adopters are implementing AIPREF to establish clear preferences and demonstrate good faith participation in developing responsible AI practices. Support from major AI systems and platforms is expected to increase as the standard matures.
Ready to implement AI preferences on your website? Use our free generator to create AIPREF-compliant configurations in minutes.
Go to Generator →