Google plans to support shared AI training opt-out via robots.txt (developers.google.com)

Google has announced plans to enhance AI training control for website owners by incorporating shared preferences through the robots.txt file. This new functionality, developed in collaboration with the AIPREF working group, allows webmasters to express their AI training preferences with a single policy token that can apply across multiple AI providers. By opting out of AI training through this streamlined approach, website owners can manage their data more effectively and reduce the administrative burden of issuing separate instructions to different AI developers. This new mechanism will coexist with existing company-specific controls, enabling a customized approach to AI interactions. The significance of this initiative lies in its potential to unify and simplify the way website owners communicate their preferences in an increasingly AI-driven landscape. It responds to widespread calls from publishers and regulators for a standardized, industry-wide solution to AI training consent, aligning with the EU’s AI Act provisions on machine-readable preferences. As more AI providers adopt this mechanism, it is expected to foster a collaborative framework for AI training governance, enhancing data privacy and control for publishers. Google aims to roll out this shared control in 2027, positioning it as a significant step forward in responsible AI practices.