---
title: Websites can prevent their content from being used for AI training
url: https://www.dataloco.com/en/websites-can-prevent-their-content-from-being-used-for-ai-training
published: 2026-11-03T20:20:39+00:00
language: en
section: AI
source: https://www.bright.nl/nieuws/2153346/zo-kunnen-websites-voorkomen-dat-hun-teksten-gebruikt-worden-voor-ai-training.html
organizations: Google, OpenAI, Anthropic, Microsoft, IETF
publisher: Dataloco
---

# Websites can prevent their content from being used for AI training

Google, OpenAI, Anthropic, and Microsoft will support a new rule allowing website administrators to indicate that their texts and images may not be used to train AI models. Therefore, websites will no longer need to specify to each AI company that their content should not be collected for that purpose separately.

Website administrators must add the new rule to their website's robots.txt file. This file allows them to include instructions for automated programs that visit their website.

User-Agent: __usage_ai-training

The proposal comes from Kevin Kelley, the technical lead for standards at Anthropic, and is being discussed within the Internet Engineering Task Force (IETF), an organization that develops internet standards.

The new rule specifically blocks the collection of information for AI training. Websites can still be visited, and their pages will continue to appear in Google search results. Administrators can also set exceptions so that certain pages may be used for AI training.

The measure does not work retroactively. If an AI company has already collected texts or images, those will not be removed by the new rule. The guideline aims to prevent such content from being collected for AI training in the future.

Anthropic currently supports the rule. Microsoft aims to do so starting February 1, 2027. Google expects to follow in early 2027, while OpenAI plans to implement the guideline in the coming months.