OpenAI will respect website owners' choice on AI training

OpenAI says it will respect a shared signal that lets website owners state whether their content can be used to train AI. The signal lives in robots.txt, a small file most websites already have. For site owners, that could mean one familiar file controls their AI training preference.

What changed

A robots.txt file is a plain text note on a website. It tells automated visitors, such as search crawlers, what they may and may not do. OpenAI will now honor a shared signal placed in that file. The signal is meant to help site operators communicate their AI training preferences to participating providers.

Who it affects

The main audience is anyone who runs a website, from a small shop to a publisher. They can say whether their content may be used to train AI. The signal is shared, so it is not tied to one company. Only providers that join are expected to follow it.

What is not known yet

• Which other AI providers have joined the shared signal.
• The exact wording or format site owners need to add to the file.
• When OpenAI will begin honoring the signal.
• How a provider's compliance will be checked.

What to do

Site owners can check whether their site has a robots.txt file and who manages it. Once the details are published, the preference can be set there. Until then, there is nothing new to change. A provider that has not joined is not expected to follow the signal.