A New AI Header: The IETF Is Building a Standard to Label AI-Generated Content

In response to the growing prevalence of AI-generated content, the Internet Engineering Task Force (IETF) has introduced a new draft proposal aimed at adding declarations within web response headers to indicate whether certain material has been created by artificial intelligence.

According to the IETF, this header is designed to be compatible with HTTP structured field syntax, providing metadata to user agents, bots (such as AI crawlers), and archiving systems (such as the Internet Archive). These systems can then determine, based on their own requirements, whether to adopt or disregard AI-generated content.

At least for now, much AI-created material remains prone to low quality and factual inaccuracies, with models often producing hallucinations or even fabricating entirely nonexistent information.

AI user agents, for instance, may fetch content from the internet and generate responses on behalf of users. If that source material itself originates from an AI model and is false or misleading, the generated responses will also be erroneous or unreliable.

Archival systems such as the Internet Archive may also choose not to preserve such content. Since AI-generated material can be produced in vast volumes with relative ease—and may be inaccurate—excluding it reduces unnecessary storage and crawling pressure on archival infrastructure.

To address this, websites that publish AI-generated material would be able to include a declaration within their headers. When AI agents or crawlers encounter this declaration, they will recognize the content’s artificial origin and may opt to ignore it, thereby avoiding errors.

The proposed syntax centers on an AI-Disclosure header, capable of declaring the generation mode, model name, provider, review team, timestamp, and other identifying details—helping automated systems quickly assess the source.

This AI-Disclosure header is entirely voluntary: websites may choose to set it or omit it at will. A header indicates that the content is AI-generated, but its absence does not guarantee that the material was created by humans.

At present, the proposal remains in draft form, with an expiration date of November 1, 2025. Should subsequent industry discussions proceed without objection, the draft is expected to be standardized—at which point all websites would have the option to adopt the AI-Disclosure header.

Related Posts:

• Proposed US Ban on Chinese Tech Impacts Autonomous Vehicles
• Google Releases Major Update and Open Source Component for Material Design Language
• Beyond HTML: The Hidden Danger of Phishing in HTTP Response Headers
• Google Antitrust: Mozilla Warns of Browser Choice Collapse