MUMBAI: The AI answer may be free for the user, but Google is testing whether the content behind it should come with a bill. The search giant is piloting a pay-per-use programme that compensates publishers when their content makes a significant contribution to AI-generated responses.
Called the AI Contribution Pilot, the initiative is being administered through Google Search Console and is currently available to a select group of publishers. Participants can track accrued earnings and payments through a dedicated reporting interface and can opt in or out of the programme.
The pilot covers content that materially influences the generation of AI responses across the Gemini app, AI Overviews and AI Mode in Google Search. But Google is drawing a line between content that helps create an answer and content that merely appears after the answer is formed.
A webpage used to confirm a fact or simply linked within an AI-generated response does not qualify for an AI contribution payment. The content must have a material role in generating the response itself.
That distinction puts a price tag, albeit a still-unclear one, on contribution rather than visibility. Unlike the large fixed-value licensing deals some AI companies have struck with publishers, Google’s pilot does not involve an upfront fee. Instead, payments are tied to qualifying usage, with publishers shown a monthly earnings figure through Search Console.
Google has not disclosed the detailed methodology used to calculate individual payouts, leaving the crucial question of how much a piece of content is worth to an AI answer still unanswered.
The company has approached at least dozens of publishers for the pilot, according to industry sources cited in the information provided. The initiative appears to have generated greater interest among small and mid-sized publishers than larger organisations, although some participants have reported that initial payments are relatively small compared with their advertising revenues.
At least one publishing executive has said negotiations with Google are continuing because the amounts offered did not meet the company’s participation threshold.
For smaller publishers, however, the proposition could be significant even if the cheques start small. The model offers a way to monetise content used by AI systems without requiring publishers to negotiate an individual licensing agreement with Google. For Google, meanwhile, it provides a live test of whether content contribution can be measured and compensated across AI products operating at scale.
Google has described the initiative as an early-stage learning pilot designed to test ways of rewarding high-quality content that helps keep generative AI responses fresh and accurate.
The experiment comes as Google’s AI products increasingly sit between publishers and their audiences. AI Overviews and AI Mode can synthesise information from multiple web sources into direct answers, while Gemini can also draw on web content when generating responses.
That is changing the old search bargain. Traditionally, publishers created content, Google surfaced links and publishers tried to turn the resulting clicks into advertising or subscription revenue. With AI-generated answers, users can increasingly get what they need without making the extra trip to a publisher’s website.
The pilot therefore introduces an important new distinction: being seen is not necessarily the same as contributing.
Google already gives publishers visibility into how their content performs across Search and Discovery. The AI Contribution Pilot goes a step further by testing a payment mechanism for material that Google determines has actually influenced the creation of an AI answer.
The move comes as AI companies face growing pressure to address how copyrighted publisher material is used. While some companies have pursued formal licensing agreements with negotiated or upfront payments, Google’s approach is instead testing a usage-linked model across its own AI ecosystem.
Google has previously experimented with AI-related arrangements involving news publishers and has run an AI feature pilot with major publishers. Its broader news AI scheme involves more than 200 titles globally, while Google News Showcase’s licensing programme covers more than 2,800 publications across 33 countries.
The company also introduced an AI Performance Report in Search Console in August, giving publishers greater visibility into how their content performs across Google’s AI-driven search surfaces. Some publishers have reported that this AI-related exposure is appearing alongside changes in traffic from traditional Search and Discover.
For now, the AI Contribution Pilot remains exactly that, a pilot. The eventual scale of participation, payment formula and economics are still unclear. But the experiment signals a potentially important change in the publisher-AI equation: content could increasingly be valued not just for the readers or clicks it attracts, but for the answers it helps machines create.
And as search moves from a page of blue links towards generated answers, that could prove to be a rather consequential change in the way the web gets paid.