Social Rocket
Google is testing usage-based payments for publishers whose content contributes to AI-generated answers. According to research by Digiday, the program runs through Google Search Console: participating publishers see an additional area there called "AI earnings", which shows monthly revenue from the use of their content.
The model is notable because Google is no longer offering only traffic or visibility in return for publisher content; in this pilot, concrete use of content in generative answers is linked to a payment. Google confirms that this is an early learning experiment. How exactly individual usage events are evaluated and paid is not yet disclosed by the company.
How the "AI contribution" pilot works
According to Digiday, publishers are paid when their content contributes to generating or grounding answers in Gemini, AI Overviews and AI Mode. The payment depends on usage; according to the report, there is no upfront payment. Publishers can leave the pilot again.
For participants, a separate earnings area appears in Search Console with monthly values and history. What is missing is the decisive level of detail: according to Digiday, publishers do not see which individual piece of content was used how often, which product type triggered the use, or according to which formula Google prices a usage event.
This black box is currently the biggest reservation. Several publishers cited by Digiday welcome in principle that AI use is finally given financial value. Others consider the sums offered so far too low and fear that early participation could weaken their negotiating position in future licensing models.
Google had already hinted at the model in June
The new payment does not come entirely out of the blue. On 18 June, Google wrote in an official public policy post that the company was testing new models of collaboration and value exchange with websites whose content materially contributes to the timeliness and factual accuracy of generative AI answers.
Google explicitly names the grounding of generative answers and refers to further publisher programs. According to the company, the News AI pilot includes more than 200 publications. Through News Showcase, Google works with more than 2.800 publishers in 33 countries; in Europe, the company states that it pays more than 5.500 publications through existing rights agreements.
The Search Console pilot now described by Digiday goes a step further: it attempts to map value not only through a flat licensing agreement, but based on concrete AI use.
This is not the same as payment for AI training
A clear distinction is important for publishers. Google already offers Google-Extended as a technical control over whether content may be used to train future Gemini models or for certain grounding applications in Gemini Apps and the Vertex AI API.
The current pilot, by contrast, is described as payment for content that supports ongoing generative answers. It cannot be concluded from this that every use of a publisher article for model training is paid, or that all Google AI products bill under the same scheme.
This distinction matters more because publishers are simultaneously struggling with less predictable search traffic. How much the business model is shifting is already shown by the trend that publishers are buying significantly more paid search traffic, while publishers are reorganizing their organizations for a post-Google and AI search era.
Why a pay-per-use model is strategically relevant
The central question of the past two years was often: if an AI summarizes content and the user no longer needs to visit the original page, what value does the publisher still receive?
Traditionally, this value consisted of referral traffic. A pay-per-use model theoretically creates a second currency: a piece of content can have economic value even if the user does not click through to the website.
For media houses, this would be a fundamental shift from the referral model to a kind of inference or grounding licensing. However, its economic viability remains completely open. If payment per use only replaces a fraction of the lost advertising, subscription or affiliate revenue, the model does not solve the traffic problem.
This is precisely why transparency and measurability are so important. Publishers would need at least to be able to trace which content generates value in which AI products and how that value is calculated. Without this data, the Search Console figure remains more of a payout than a reliable business model.
Google is also under regulatory pressure
The pilot does not exist in a vacuum. EU competition authorities are currently questioning publishers about Google's ability to opt out of AI search separately without affecting rankings in classic search. Reuters reported in early September that the feedback is part of an ongoing EU investigation. In the UK, Google has also been asked to give publishers more control over how their content is used for AI features.
Google, for its part, emphasizes that publishers should receive more control, new Search Console metrics and additional partnership models. For publishers, these three points will likely go together in the future: control over use, traceable measurement and an economic return.
For marketers, this is also relevant. If publisher content no longer has to be optimized only for clicks, but must also function as citable and usable sources for AI systems, the content economy shifts. Visibility in generative answers then becomes not only an SEO topic, but potentially also its own revenue and licensing channel.
Still a pilot, not a new standard
The most important caveat therefore remains the status of the program. Google speaks of an early learning experiment; Digiday reports a growing but limited group of publishers being approached. There is so far no general availability, fixed price lists or publicly traceable billing rules.
Nevertheless, the step is relevant: Google is beginning to treat individual AI usage events as a payable value. Whether this becomes a viable market for publisher content or merely a small additional fee alongside continuing declines in referrals now depends above all on price, transparency and bargaining power.