SPUR publishes a standard to track AI use of news content

The SPUR Coalition has published a Content Telemetry Standard designed to show publishers when AI systems retrieve, use, cite, display or engage with their journalism. The framework seeks to establish a shared reporting system for AI companies and a clearer record of how news content contributes to AI answers. The first version follows a draft released for public comment in June. SPUR’s proposed telemetry profile would send event-level reports in real time to an endpoint selected by the content owner, while the standard and profile are intended to protect user privacy and intellectual property. It defines five stages of content use: retrieval, grounding as evidence or reference for an AI response, citation, display to a user and engagement. Engagement can include actions such as clicking, copying, sharing or directing an AI agent to act. The system is meant to give publishers data they have often lacked as chatbots and AI-enabled search products draw on web material to answer users. It could also provide a common basis for licensing and compensation frameworks, as well as evaluation and auditing of AI responses’ quality, integrity and usefulness. SPUR is a nonprofit publishers’ coalition established earlier this year to shape the technical and commercial environment for generative AI’s use of intellectual property. Its founding members include the BBC, Financial Times, Guardian Media Group, Mediahuis, Sky News and Telegraph Media Group; it subsequently announced the addition of 30 members at WAN-IFRA’s World News Media Congress in Marseille. The coalition says the specifications are being developed with publishers, website and content owners, analytics and reporting platforms, standards bodies, content marketplaces, frontier-model labs and developers. Its founding open letter committed the group to creating common industry standards and responsible routes for original journalism to be used sustainably. SPUR has invited OpenAI, Anthropic, Google, Meta and Microsoft to help shape the standard’s implementation, according to Digiday. Companies accepting the invitation could join an invitation-only AI Licensing Advisory Board, expected to have about 20 members and to hold its first meeting in October. The published version added support for multimodal material rather than text alone and was refined to ease integration with C2PA and other provenance standards. SPUR is also working on implementation and proofs of concept, with pilot programmes planned, and is developing agent tooling intended to make adoption and reporting easier for AI companies. The initiative comes amid publishers’ concerns that AI products can answer users without directing them to the underlying articles, potentially reducing traffic and advertising revenue. The Reuters Institute found that one-third of users consistently click source links in AI-generated search answers, while 28% rarely or never do so. Its account of SPUR said the coalition aims to help members secure fair value for journalism while leaving them free to negotiate licensing agreements individually.