TREN
← News and archive

Website content and AI training preferences · Announced

Should your content train AI? A shared preference signal for website owners

Published: 06/11/2026 · Event date: 30/10/2026

What happened?

Wanting a business to be found online does not mean making the same choice about every use of its content. On 30 October, OpenAI announced support for a shared robots.txt signal to communicate AI training preferences to multiple participating providers. Implementation is planned for the coming months, while existing crawler controls remain in place.

How could this help me?

Publishers and website managers can start by treating search visibility and model training as separate decisions. OpenAI’s current documentation distinguishes OAI-SearchBot for search from GPTBot for training preferences. A shared signal could reduce the work of communicating the same choice to multiple providers. It does not automatically change the behaviour of every provider.

ExampleIllustrative scenario

A consultancy wants prospective clients to find its blog, but chooses a different policy for model training. The website owner and content team first write down those two decisions separately.

Read the full exampleShow less

A developer compares the current robots.txt file with OpenAI’s crawler documentation and explains which rule serves which purpose, rather than adding one blanket block. They check the proposed change in a test environment to avoid unintentionally blocking search access. When implementation details for the shared signal become available, they review the policy again. The aim is to avoid confusing discoverability with training preferences; the example promises neither visitor growth nor guaranteed protection of content.

Sources