According to a report from The Information, with Digiday having first reported on the program, Google has been running a pilot that compensates publishers for what they contribute to its AI-powered search features. The Verge's account of that reporting says the test began less than a year ago and that roughly 100 publishers are involved.
Rather than a flat licensing fee, the mechanism described is contribution-based: participating publishers are reportedly paid according to the extent of their content's contribution to AI Overviews and AI Mode in Search, as well as to the Gemini chatbot. For anyone whose income depends on search referral traffic, that framing matters, because payment is tied to use inside AI answers rather than to clicks on a link.
Payouts reportedly vary widely by tenure in the program. One publisher that joined when the pilot first started earned more than $1 million over a year, according to The Information, while roughly $50,000 to $60,000 has been earned by another that joined a few months ago. Those two figures are the only scale indicators in the supplied evidence, and they describe individual participants, not an average or a published rate card.
Against a backdrop of publisher complaints and litigation over Google's AI-focused search changes, and of regulatory attention, the program arrives. In June, The Verge notes, the UK ruled that appearing in AI search features must be something publishers are allowed by Google to opt out of, while an investigation into the impact on web traffic was opened by the European Union, which recently ordered changes to its search engine.
For freelancers, designers and developers, the practical read is indirect but real. If compensation for AI-surface usage becomes a normal part of publisher economics, the outlets that commission freelance work may gain a revenue line that is decoupled from pageviews — but the evidence here covers only a reported pilot with about 100 participants, and no terms, eligibility rules or rates have been published.
The evidence does not say how Google measures a given publisher's contribution, whether payments are recurring, whether the program will expand beyond the reported cohort, or whether participants must accept any conditions in exchange. It also does not establish that the program is a response to the UK ruling or the EU investigation; the two developments are reported alongside each other, not causally linked.
One structural detail is worth separating from the headline. The reported payments cover three distinct surfaces — AI Overviews, AI Mode in Search, and Gemini — and the evidence does not break out how compensation differs across them. A publisher whose material is heavily used in one surface but not the others would, on the described logic, be paid differently from one used evenly across all three.
The Verge's piece also carries a set of unrelated headlines in its feed, including items about OpenAI's Dots, an Apple device, car data privacy, OpenAI DevDay 2026, a Firefox redesign, and a native ad. None of those items is part of the Google pilot story and none should be treated as supporting detail for it.
What remains unknown is substantial: the identity of participating publishers, the formula behind the payments, whether the program is contractual or discretionary, and whether Google has confirmed any of it. The Verge attributes the program to reporting by The Information and Digiday rather than to a Google announcement, so the entire account should be read as reported, not confirmed.
For this audience, the honest implication is to watch publisher-side revenue experiments rather than to plan around them. Freelancers negotiating with outlets that rely on search traffic have no published rate or eligibility information to cite, and the reported earnings figures describe two unnamed participants at different points in a sub-year-old test.
A reasonable conclusion is that the reported pilot is a signal about how AI search usage might eventually be compensated, not yet a program anyone can join or budget against. Until Google confirms terms or the reporting is corroborated by named participants, treat the $1 million and $50,000–$60,000 figures as reported anecdotes from a limited cohort rather than as evidence of a market rate.