A UK regulator now requires Google to show how it uses web content in AI search. Brands, retailers, and publishers gain new leverage-but must act to control their data.
Google must now reveal how it uses content from any website-including retailers, brands, and newsrooms-in its AI-powered search features, following a new requirement from the UK’s Competition and Markets Authority (CMA). This change means that every organization with a website is treated as a publisher, giving them new visibility and control over how their material is used by AI systems.
The CMA’s Publisher Conduct Requirement, imposed on June 3, defines a publisher as any party making content available online. This includes not just traditional news outlets, but also e-commerce sites, software companies, and direct-to-consumer brands. For AI agents, there is no distinction between a months-long investigation and a product page-both are simply data sources for generating answers.
Under the new rules, Google must clearly attribute content used in AI search results and provide users with a direct path back to the original source. When Google’s AI answers questions like “which running shoe” or “which accounting tool” using your product pages or reviews, it must account for that use and link back to your site. However, this control is opt-out: unless you actively withhold your content, it remains available for Google’s AI features, either page by page or across entire site sections.
This shift affects not only major publishers but also mid-sized brands and small content creators. AI companies have acknowledged that their models depend on the quality of the data they ingest, and that includes content from across the web. As OpenAI has stated, training leading AI models would be impossible without using copyrighted materials. This dependency gives all content owners leverage, regardless of size.
While large organizations like The New York Times may have the resources to pursue lengthy legal battles over AI use of their content, most brands and smaller publishers do not. Their practical leverage now lies in monitoring how their content is used, controlling access, and demanding transparent measurement they can trust.
The CMA’s ruling also requires Google to report engagement metrics for content used in AI search, such as impressions, clicks, and click-through rates. However, consistency remains a challenge: different platforms measure these metrics in different ways. For example, Microsoft’s Clarity AI citations dashboard and Google’s upcoming Search Console update both track AI-driven engagement, but their numbers and definitions do not always align.
Without standardized measurement, it is difficult for publishers and brands to compare performance across platforms. Industry groups are now working on open standards for recording how AI systems retrieve, cite, and drive engagement with content. The CMA has indicated it may consider such standards as they emerge, and a proposed shared format is set to open for public comment on June 12.
Industry leaders argue that standardized reporting is critical as AI agents increasingly access and use web content, often without visible citations. The terms for how publishers can see and control their content in AI-generated answers are being set now, and those who participate in shaping these standards will have the most influence.
Recent data shows that direct traffic to publisher sites is declining, with social and video platforms becoming primary news sources. This shift, as reported in a recent analysis of Reuters Institute findings, highlights the growing importance of understanding and managing how content is used and attributed across digital platforms.
OpenAttribution, a neutral non-profit, is developing open standards for AI content use, attribution, and transparency. The organization aims to help publishers, brands, and creators track when their work is retrieved, cited, or engaged with by AI systems, and to provide platforms with a consistent, privacy-preserving way to report that usage.
