Technology
OpenAI Google Microsoft back AI training opt-out token
The Indian Express·
OpenAI, Google, Microsoft and Anthropic are moving to adopt a shared policy token that lets publishers signal content is barred from AI training. Anthropic's Kevin Kelley proposed the Internet Engineering Task Force standard for robots.txt. Anthropic has already adopted the signal. OpenAI said it will implement it in coming months. Google and Microsoft plan support next year.
Prism
What It Means For You
- Publishers could place one policy token in robots.txt instead of separate rules for each AI crawler.
- Sites that adopt the token would tell major AI crawlers to skip AI training use.
- Search visibility rules remain separate from the training opt-out signal under existing Google and Microsoft controls.
What's Happening
- The shared signal was proposed inside the Internet Engineering Task Force process by Anthropic engineer Kevin Kelley.
- Anthropic said it already honours the token; OpenAI pointed to coming months for implementation.
- Google and Microsoft said they plan support next year, while Meta has not given a date.
Why a Shared Signal
- Publishers have challenged unlicensed use of their text to train generative AI systems.
- Robots.txt already governs crawlers, but preferences often had to be set per AI provider.
- Other approaches include Cloudflare's default AI-crawler blocks and Really Simple Licensing fees.
all-news