Google is testing a new payment model that could change how publishers earn from content used by artificial intelligence. The company has begun inviting selected publishers to an AI Contribution Pilot, under which they can receive payments when their content makes a significant contribution to AI-generated answers.
The programme is being run through Google Search Console and covers content used across the Gemini app, AI Overviews and AI Mode. Rather than paying publishers a fixed licensing fee, Google is testing a model based on the value of content used to generate an answer.
The pilot is currently invitation-only. Industry reports indicate that at least dozens of publishers have been approached, with the programme appearing to attract greater interest among small and mid-sized publishers than large media organisations. It has also expanded beyond traditional news publishers, suggesting that Google is testing whether the model could work across different types of websites and content.
Publishers that join the programme get an additional AI Contribution section inside Search Console. The dashboard shows a monthly earnings figure along with some historical data. However, it does not currently explain in detail how Google arrived at that amount, leaving publishers with little visibility into the calculation behind their payments.
That lack of transparency is one of the biggest questions surrounding the experiment. Google is not simply counting how many times a website’s content is accessed by its AI systems. Instead, the company is assessing whether the material significantly helped generate the answer.
This means a publisher is not automatically paid simply because its website is cited or linked in an AI response. Google distinguishes between content that helps shape an answer during its generation and information that is used later to confirm a fact or provide a citation. Only the former qualifies under the pilot.
The distinction is important because Google’s search experience is changing rapidly. Traditional Google Search directed users towards individual websites, allowing publishers to earn through advertising, subscriptions, ecommerce or other forms of audience engagement. AI search increasingly gives users a direct answer without requiring them to open several links.
That shift has created a difficult problem for publishers. Producing original journalism, research and specialist information requires money and people, while AI systems can use that work to create concise answers for users. If readers get what they need directly from an AI response, publishers could see fewer visits and less advertising revenue.
Google’s new approach attempts to address part of that problem by putting a monetary value on content that contributes to an AI answer. The company has described the initiative as an early-stage learning pilot aimed at finding ways to reward high-quality content while continuing to provide websites with traffic and other tools.
Google had already indicated in June that it was testing new ways to work with websites whose content helps keep its generative AI responses fresh and accurate. The company referred to the use of web content for “grounding”, a process in which an AI system consults external information while producing an answer. The latest pilot appears to put a payment mechanism around that relationship.
The model is also different from the large, negotiated AI licensing agreements signed by some publishers with companies such as OpenAI. Google’s pilot does not involve a publicly announced fixed payment or upfront licensing fee. Instead, earnings accumulate based on qualifying contributions, and publishers can leave the programme at any time.
Early reactions from publishers have been mixed. Some participants see value in being part of an experiment that could eventually create a new revenue stream. They also see Google’s willingness to make direct payments as an important precedent at a time when traditional referral traffic from search is under pressure.
Others have questioned whether the money is meaningful enough. Industry reports indicate that some early participants consider the returns small compared with the advertising revenue their websites generate. The limited information about how payments are calculated has added to those concerns.
The timing of the pilot is significant. Google has been expanding its AI search products while giving publishers more tools to measure how their websites appear in generative AI results. Since August 31, its Search Console has also provided broader reporting on a site’s visibility across Google’s AI search features.
For publishers, the bigger issue is not just the amount of money being offered today. The pilot could help establish a new economic relationship between Google AI and online content, where websites are compensated for contributing information to an answer rather than only for sending a user to a webpage.
Much will depend on whether Google expands the programme, increases transparency and develops a clearer method for valuing individual pieces of content. If the experiment grows, it could influence how publishers negotiate with AI companies and how the wider digital media industry approaches AI content licensing.
