Google's experimental program for paying publishers whose content feeds AI answers has around 100 participants, and for many of them the payouts amount to less than one-tenth of 1% of their advertising revenue. That is the core finding of reporting published by The Information on September 30, 2026, and it is the first concrete look at the economics of the AI contribution pilot — the Search Console-based scheme Google has been quietly testing since late 2025.
The significance goes beyond the small dollar amounts. This is the first time Google has ever put a price on publisher content inside its own AI answers, and the price it has set so far is close to zero for most participants. For an industry already squeezed by AI Overviews and AI Mode reducing clicks, the numbers will land as evidence that voluntary, Google-set compensation may not arrive at meaningful scale — or at all.
What happened
The AI contribution pilot pays publishers when their content "significantly contributes" to AI-generated responses in three Google surfaces: AI Overviews, AI Mode, and the Gemini app. It runs inside Search Console. Publishers who opt in see an "AI contribution" panel with a monthly earnings figure and some history — and, per multiple participants, little else. There is no breakdown of which pages earned what, which queries triggered the contribution, or how Google weighs one source against another.
According to The Information's report:
- Payments vary enormously. One publisher that joined a few months ago has earned about $50,000 to $60,000 so far — real money, but a minor share of its business. Another early participant is earning more than $1 million a year, a much larger slice of its revenue. Smaller sites have earned less than $1,000 over several months.
- For several small and midsize blogs and websites, the payments amount to less than 0.1% of their advertising revenue.
- Payments fluctuate month to month without explanation. Some participants told The Information they have no idea how Google calculates the amounts.
- One publisher described the arrangement bluntly: Google uses publishers' content and decides for itself what that content is worth.
The program is invitation-based. Digiday, which first documented the pay-per-value structure in detail on September 14, reported that dozens of publishers had been approached, that participation skews toward small and mid-sized publishers, and that it has widened well beyond news. Google has confirmed the scheme is an early-stage learning pilot for testing how to reward high-quality content, on top of the traffic and tools it already provides, and says publishers can opt out at any time.
The timing of the numbers is notable. The Information's figures arrived two weeks after Digiday's structural reporting and roughly a month after Google began rolling out the AI performance report in Search Console, which shows how often pages appear in AI-generated answers. Until now, the public debate about AI compensation has been conducted mostly in hypotheticals. These are the first actual payout figures to enter that debate.
Why it matters
The pilot's design is a departure from how AI companies have paid for content so far. OpenAI struck lump-sum licensing deals — News Corp's was reported at roughly $250 million over five years. Reddit disclosed more than $200 million in data licensing agreements in its IPO filings. Anthropic, having largely avoided publisher licensing, settled its authors' copyright litigation for $1.5 billion. Google, by contrast, has avoided the lump-sum model entirely and instead set up a contribution-based payment that it alone calculates.
Three structural problems follow from that design, and the new numbers make all three concrete.
First, the pricing is unilateral and opaque. Google defines "significant contribution," computes the value, and discloses only a monthly total. That means publishers cannot audit the number, predict it, or negotiate it in any conventional sense. One executive with knowledge of the program called it "quite black box," according to Digiday. When the party that uses the content is also the party that prices it, the resulting figure reflects Google's interests and budget — not a market price.
Second, the amounts are small relative to the value being extracted. A payout below 0.1% of ad revenue is not, for most publishers, a revenue stream; it is a rounding error. If AI answers increasingly satisfy queries without clicks — and Google's own AI performance report is now making that displacement measurable inside Search Console — then the compensation question is whether direct payments can offset lost referral traffic. At the reported levels, they cannot.
Third, early acceptance may set the anchor. Several publishers are already inside the tent, attending weekly calls with Google that participants describe as collaborative. Publishers who accept these initial rates may find it harder to argue for higher compensation later, because Google can point to the pilot and say it already pays for AI use. Industry executives have made exactly this argument: accepting low payments now could weaken publishers' collective negotiating position, effectively handing Google the ability to claim it has resolved the compensation issue at minimal cost. This concern sits against a regulatory backdrop where publishers have also sought the right to opt out of AI use without losing search ranking — a right that has featured in antitrust discussions around Google's AI products.
There is a counterargument, and it deserves a fair hearing. Some participating publishers argue that being inside the pilot beats standing outside hoping the old referral economics return. One exec told Digiday: "Do I wish they were more transparent? Definitely. But I'm hopeful that the fact that they're setting a precedent for exploring paying publishers directly for content through this is meaningful." Another framed the pilot as an early test of a marketplace for inference data — a system where publishers could learn, the way they learned from Search Console's performance reports, what content is valuable to AI systems as well as to readers. That feedback loop does not exist anywhere else yet.
Both things can be true. The precedent matters, and the price discovered so far is close to zero.
What it means for SEO and GEO practitioners
If you work in SEO or GEO, this story is easy to misread as publisher-side industry news that does not touch your work. It is not. The pilot's payment logic tells you exactly what Google's AI systems consider valuable at the moment they compose an answer — and that definition now has money attached to it.
The payment trigger is generation-phase contribution, not citation. Google's help documentation for the pilot states that it pays when content contributes significantly to the creation of an AI-generated response — the phase where the web content is actually used to compose the answer. Content that is merely consulted to verify a fact, or that appears as a link after the answer is generated, does not qualify. Read that as a ranking system for source influence, not a citation count. A page that shapes the wording, data points, or structure of an AI answer is doing something different from a page that merely appears in a source list, and Google is now distinguishing between the two with dollars.
"Black box" cuts both ways. Publishers cannot see the calculation, but that also means nobody can game it directly. What survives as durable strategy is the boring stuff: original data, up-to-date facts, clear definitions, first-hand experience — the inputs that AI answers actually need to draw on. The Information reported that some publishers in the pilot see it as a chance to learn what content Google's AI values and adapt theirs to match. That adaptation instinct is right, but the same qualities that make a page quotable also make it rankable and citable, so the work is not wasted if the pilot never scales.
The AI performance report is your closest proxy. Most sites have no access to the contribution panel. But the AI performance report in Search Console — which Google began rolling out in August 2026 — shows impressions and visibility across AI surfaces for everyone. Track which pages appear there, treat those pages as your closest observable signal for what Google's AI surfaces rely on, and maintain their accuracy and freshness. Discrepancies between pages that get AI impressions and pages that get clicks are now measurable, and they quantify your own AI-driven traffic displacement.
Assume payment-based revenue will not save publishers soon. If you run or advise a content business, do not build a financial model that assumes AI-era direct payments arrive at meaningful scale within the next year. The reported figures — ranging from under $1,000 over several months for small sites to $1 million-plus annually for an early large participant — suggest payments will remain concentrated among a small invited group for the foreseeable future, with amounts that most mid-tier publishers cannot rely on.
What to do now
- Check Search Console for the AI contribution panel and the AI performance report. The contribution panel appears only for invited participants; the AI performance report is rolling out broadly. If you see either, start recording the numbers monthly. Baseline data from the pilot period will be valuable if the program later scales or if terms change.
- Audit your AI-visible pages for generation-phase value. Take the pages that appear in AI performance reports and ask whether they contain original data, named specifics, dated facts, and definitions an AI would need to borrow to compose an answer — or whether they mostly restate what other pages already say. Prioritize the former in your refresh cycles.
- Do not accept the first price as the market price. If your organization is approached for the pilot, ask in writing how contributions are valued, request historical breakdowns, and coordinate with peer publishers before accepting. The payout figures now public are the opening position of a single party in the transaction.
- Watch whether the pilot's definition of "significant contribution" gets documented further. Google's current help page describes the eligible generation phase but discloses no calculation method. If Google ever publishes contribution-level reporting — the feedback loop pilot participants are asking for — it would become the most important dataset in GEO, and the first to know about it will be those already tracking AI surfaces closely.
The numbers from The Information settle one question and open another. Google is willing to pay some publishers for AI use — but so far, for most, barely. What remains unresolved is whether that is the starting point of a real market for content in AI answers, or the ceiling.
Sources
- The Information — Google Is Paying About 100 Digital Publishers for AI Overviews: https://www.theinformation.com/articles/google-paying-100-digital-publishers-ai-overviews
- Digiday — Google rolls out pay-per-value AI licensing program to publishers: https://digiday.com/media/google-rolls-out-pay-value-ai-licensing-program-to-publishers/
- Google Search Console Help — AI contribution pilot (help document, screenshotted in reporting): https://support.google.com/webmasters/answer/17257106
- Open Society Foundations — Publishers Won the Right to Say No to Google's AI: https://www.opensocietyfoundations.org/voices/publishers-won-the-right-to-say-no-to-google-s-ai-here-s-what-it-means
- LLM Pulse — Every AI Content Licensing Deal, Mapped (2023–2026): https://llmpulse.ai/blog/ai-content-licensing-deals/
