OpenAI, Google, Microsoft, and Anthropic have backed a shared standard allowing websites to opt out of generative AI training through a single setting.
OpenAI, Google, Microsoft, and Anthropic have backed a shared standard allowing websites to opt out of generative AI training through a single setting.
The approach uses a common __usage_ai-training signal in robots.txt, reducing the need for publishers to configure separate restrictions for each AI provider. Anthropic has begun implementation, while the other companies plan to roll out support over the coming months and into 2027.