Google Will Pay Publishers for AI Use. Should You Opt In or Block?

Google is testing a program that pays some publishers when their content significantly contributes to answers in AI Overviews, AI Mode and the Gemini app. Google confirmed to Digiday that this "AI Contribution" pilot is an early-stage learning program, and participants see their earnings in a new Search Console panel (Search Engine Journal, 14 September 2026). It is invite-only, the payment formula is not public, and the reported payout figures come from press sources, not Google.

So for almost every site, the real choice today is not "opt in or block." It is three separate decisions that are easy to confuse. This piece explains what is known, what is not, and how to make each call.

What is Google's AI Contribution pilot?

Here is what the reporting, and the Search Console help text quoted in it, describe.

What it pays for. Content that "contributes significantly" to an AI answer in the Gemini app, AI Overviews or AI Mode. Only use during answer generation counts. Pages that are linked or used to confirm facts after the answer is written "don't qualify for AI contribution," according to the help text quoted by Digiday and Search Engine Journal.

Where it shows up. Participants get a Search Console panel with monthly earnings and some history. The help text notes that earnings and payments are tracked separately, and the amount paid can differ from reported earnings because of tax deductions or local payment thresholds.

Who is in it. Invited publishers only. Digiday reported that at least dozens of publishers had been approached but could not confirm the total. Search Engine Roundtable, citing The Information, reported about 100 participating publishers (30 September 2026).

How to leave. Participants can leave at any time from Search Console settings.

What Google has confirmed. That it is an early-stage pilot meant to reward high-quality content alongside the traffic and tools Google already provides. Google pointed Digiday to an earlier post about grounding that did not name the pilot or describe payment terms. We found no Google announcement or public help page describing rates.

How much does it pay?

Nobody outside the pilot knows for sure, and Google has not published rates. Treat every number below as reported, not confirmed.

Search Engine Roundtable, citing The Information (30 September 2026), reported that payments to several small and mid-sized sites were less than one-tenth of 1% of their advertising revenue. The same report described wide variation: one publisher that joined a few months earlier had earned roughly $50,000 to $60,000, smaller sites had earned under $1,000 over several months, and one early participant was earning more than $1 million a year. An executive quoted by Digiday called the panel "quite black box," because it does not show how earnings are calculated or which pages contributed.

The honest summary: for most publishers, this is not yet a meaningful revenue line, and there is no way to forecast what it would pay you.

Why is Google doing this now?

The pilot lands in the middle of a wider fight over who gets paid when AI uses content. Publishers are suing AI companies, including USA Today Co, which sued OpenAI on 9 October 2026 seeking $250 million over articles from 19 titles (Press Gazette). Licensing deals keep arriving too: OpenAI signed the Times of India group and Indian Express on 23 September 2026 (Press Gazette deals tracker). And Google has already given site owners a Search Console control to keep their pages out of AI features, which we covered in our piece on Google's AI opt-out.

What are the three decisions, and how do they differ?

This is where most confusion sits. There are three separate controls, and they do different things.

1. The AI Contribution pilot (money). Joining or leaving decides whether you are paid for contribution. It is invite-only, so most sites cannot opt in. Nothing in the reporting suggests that leaving the pilot removes your content from AI features. Refusing the money and refusing the use are different decisions.

2. The Search Console AI opt-out (use in Search AI). This domain-level setting stops Google using your pages in AI Overviews, AI Mode and AI Overviews in Discover, while your site stays indexed and ranked. It launched in the UK first in June 2026 and reached all sites on 31 August 2026, according to Google's help documentation as reported by Search Engine Journal and Search Engine Land. Page-level control is planned for March 2027. Our Google AI opt-out guide covers who should use it.

3. Google-Extended in robots.txt (Gemini training and grounding). Google's crawler documentation describes Google-Extended as a standalone token that controls whether crawled content is used to train future Gemini models and for grounding in Gemini Apps and Vertex AI. Google states it does not affect a site's inclusion in Google Search and is not a ranking signal.

The interaction worth noticing: the pilot pays for contribution to the Gemini app as well as Search. If you block Google-Extended, your content is not used for grounding in Gemini Apps, so you may also be removing one of the surfaces the pilot pays for. Google has not said how the two interact, so test before assuming.

Should you opt in or block?

Start from what each surface is worth to you today, not from a payout nobody can quote.

If you sell products or services. AI visibility is a marketing channel. Being cited and named in AI Overviews, AI Mode and Gemini is the goal. Blocking would remove you from answers your buyers are reading. The pilot's money, if you are invited, is a bonus, not a reason.

If you are a publisher whose business is the content itself. The calculation is harder. AI answers can satisfy a reader without a click, and the reported payments are small. Opting out of Search AI features may be leverage, but it also removes you from a growing surface. If you are invited to the pilot, joining costs you nothing you were not already giving. Track what it actually pays for a few months before deciding anything bigger.

If you are unsure, measure first. Three numbers make the decision:

• How often your pages appear in AI features. The Search Console generative AI report shows impressions, though not clicks.
• How much referral traffic AI surfaces actually send you, from your analytics.
• What the generated answers say about your brand and which sources they credit.

How to set your AI access policy without guesswork

Robots.txt is now a policy document with many audiences. Google, OpenAI, Anthropic, Perplexity and others each run multiple crawlers with independent tokens, some for training, some for search, some for user-triggered fetches. Blocking the wrong one can quietly remove you from an engine you wanted to be in.

A few practical rules:

Decide per bot, not per company. A training crawler and a search crawler from the same company can be treated differently. Our comparison of llms.txt vs robots.txt vs sitemap explains which file does what.

Do not rely on llms.txt for control. It is a suggestion file, and the evidence that engines use it is thin, as we found in does llms.txt work. Robots directives are what crawlers are documented to respect.

Diploria's AI Access view shows your robots policy bot by bot, including Google-Extended and the separate crawler tokens from other AI companies, and Robots Diff flags when your robots.txt changes so a silent block does not go unnoticed. AI crawler detection and LLM Traffic (from GA4) then show whether bots are visiting and whether AI surfaces are sending people.

If you want a quick read on how visible you are in AI answers before making any access decision, the free AI visibility checker tests five engines in about 60 seconds.