Google confirmed on September 14 that it’s running an “AI contribution pilot,” a Search Console-based program that pays publishers when their pages materially shape answers inside the Gemini app, AI Overviews, and AI Mode. The pilot was first spotted in April 2026 and denied at the time; Google eventually pointed reporters back to a June 18 blog post about “partnering” with sites whose content keeps generative responses fresh and accurate through grounding.
The mechanics are unusually specific. Publishers opt in through Search Console, an AI earnings widget surfaces a monthly figure with some history attached, and payout is triggered only when a page “significantly contributes” to a generated answer. The welcome screen reads: “Start earning with AI contribution pilot. Accumulate earnings when your content contributes significantly to AI-generated responses in participating products”. Google’s help text is equally emphatic about the negative case: “Web content may also confirm facts or be linked to within Google’s AI after a response is generated, but those instances don’t qualify for AI contribution.”
Being cited isn’t the bar. Shaping the answer is.
Note the surface asymmetry. The contribution pilot spans three products; Search Console’s generative AI performance report, which rolled out to every site worldwide on August 31, only measures two of them (AI Overviews and AI Mode). Google is willing to pay for Gemini-app usage it isn’t yet willing to fully report on.
Digiday puts the invite list at “at least dozens” of publishers, skewing toward small and mid-sized sites and reaching well beyond news. That’s a striking cadence next to Google’s neighboring programs: 200-plus titles in the broader news AI scheme, and more than 2,800 publications across 33 countries inside Google News Showcase. The narrow rollout is a signal about who Google thinks is producing the source-of-truth passages its models actually lean on.
The economics are contested. One source told Digiday early returns have been “peanuts” next to ad revenue; one exec said the “lowball numbers” Google floated didn’t clear an internal threshold. An exec who did join described the model as “It’s quite black box,” but concluded it was “far better to be inside Google’s AI licensing tent” than to wait on referral traffic that may not arrive. Search Engine Journal flags a further hazard: accepting a check now could weaken a publisher’s leverage in any later negotiation, since Google can point to payment already made.
Luke Stillman, managing director of Madison and Wall, framed the pilot as “provide some incremental upside for publishers while also helping Google reduce legal and reputational risk around how content is used.”
For small operators, the read-through is more useful than the checks. Google has now defined, in production, the difference between content it links to and content it actually uses. That’s the bar to clear for discovery in Google’s AI search report, alongside the exact-match AI Mode ad surface small businesses now have to plan around and the rise of AI tools as a top local-discovery channel.
The invite list tilting toward smaller sites tells its own story. Depth is legible to the model. Scale, on its own, isn’t.
Sources
- https://digiday.com/media/google-rolls-out-pay-value-ai-licensing-program-to-publishers/
- https://searchengineland.com/google-tests-paying-publishers-for-using-its-content-in-ai-mode-ai-overviews-and-gemini-488382
- https://www.seroundtable.com/google-al-contribution-pilot-42076.html
- https://www.searchenginejournal.com/google-tests-paying-publishers-for-ai-answers-via-search-console/589414/
- https://www.storyboard18.com/digital/google-tests-pay-per-use-for-publishers-in-ai-responses-ws-l-110639.htm