Cloudflare Unveils New Policy to Let Creators Control How AI Bots Use Their Content

Cloudflare has unveiled a Content Signals Policy that lets creators specify how AI systems may access and use their content, including whether it can be used for training. The policy, which aims to be machine-readable via updated robots.txt instructions, lets website owners answer questions such as “Yes” to usage, “No” to deny it, or signal no explicit preference, and it distinguishes between use cases like search, input, and training.

Cloudflare says it will help operators update their robots.txt files and has released tools and templates to tailor “content usage rights” for individual sites.

The change could affect traffic and monetisation for publishers, as AI tools may answer user queries without visiting the source site, though sites can specify restrictions such as “AI may not train on this content.” More than 3.8 million domains currently use Cloudflare’s robots.txt management service to block AI training, with industry responses largely positive from figures including the News Media Alliance and platforms like Quora and Reddit.