Five stories published between July 23 and October 1, 2026 tell one story from five directions. Google’s AI answers are becoming the destination instead of the referral, and the fight over who supplies that content and who gets paid for it has moved out of the courtroom and into the crawler, the regulator’s questionnaire, and the licensing deal.
If you run a web design or marketing agency, these five items are not industry gossip. Each one changes something about how a client you already serve is described, ranked, or recommended by an AI system. None of them are settled.
Selection Criteria: Why These Five Stories
Before the list, here is how I picked it.
- Each story changes what an agency owner or a local business can actually rely on, not just what is interesting to read.
- Each comes from a primary report or a major outlet, and each source is linked so you can read the original yourself.
- Together they cover all three layers of this problem: the answer surface, the content supply behind it, and the legal and regulatory rules around both.
I am not going to blend them into one big prediction. I will report what each source said, separate my observations from their reporting, and name what each story does not tell us.
Quick Takeaways
- A federal judge dismissed antitrust lawsuits from Chegg and Penske Media, closing off private antitrust litigation as a publisher remedy over AI Overviews, at least on the theory the publishers argued.
- AI Overviews now appear in roughly 43% of Google searches, up from about 15% a year earlier, according to Similarweb. Semrush puts the figure near 48%.
- Publishers including Reddit, USA Today, Politico, Reuters, and the Economist are weighing whether to block Google’s crawlers or opt out of search entirely.
- EU antitrust regulators are asking publishers whether Google’s proposed AI search opt-out actually addresses competition concerns.
- Yelp is licensing its reviews and business data to OpenAI, and Yelp Reservations and Waitlist now work inside ChatGPT.
1. US Judge Dismisses Chegg and Penske Media Antitrust Lawsuits Over AI Overviews
What Happened
Judge Amit Mehta dismissed separate antitrust lawsuits from Chegg and Penske Media, both of which alleged that Google unlawfully used their content in AI Overviews while draining traffic away from their sites.
The publishers’ core theory was reciprocal dealing, the argument that by using their content in AI Overviews, Google created an expectation of something in return. The court did not accept it. An expectation is not an agreement.
Reuters describes the outcome as a landmark win for Google and a closing off of private antitrust litigation as a publisher remedy.
Why It Matters
For publishers, one route to leverage just closed. The legal system is not going to restore the old balance of referral traffic on a schedule anyone can plan around.
For agency owners, that is the useful takeaway. If you are waiting for a court to fix the referral economy before you adjust how your clients show up in answers, you are waiting on the wrong clock. The second thing this ruling clarifies is where the remaining leverage sits. It sits in opting out and in regulators, which are the next two stories.
The Limit and the Next Step
Where this stops. A dismissal is not a ruling that using publisher content is fair, harmless, or good for the open web. It is one legal theory failing. Dismissals can be appealed, and complaints can be refiled on different grounds.
Next step. Read the reasoning if you advise publishers. Otherwise treat the ruling as a signal about timing, not about outcome.
Source: Reuters
2. AI Overviews Now Appear in Roughly 43% of Google Searches
What Happened
New Similarweb data reported by TechCrunch shows AI Overviews appearing in about 43% of Google searches, up from roughly 15% a year earlier. Semrush puts the number near 48%. AI Mode visits jumped from 126 million to 279 million over the same period.
TechCrunch frames it as a fundamental shift. Google is no longer primarily a list of blue links. Google itself is the destination.
Why It Matters
This is the number to carry into client conversations this quarter, because it is the clearest available measure of how fast the surface has changed. Two measurement firms working independently land in the same range: a low-40s to high-40s share, up from the mid-teens twelve months earlier.
Practically, it means the unit of competition changed. Ranking well on a page of blue links is no longer the whole job. What gets said inside the answer, and which business gets named in it, is now part of the same job.
For local businesses, the inputs that appear to travel into these answers are the boring ones. Consistent business details. A current and populated review profile. Content that actually answers the question a person is asking. That is SEO fundamentals plus genuinely useful content, applied to a different surface.
Where this stops. The percentage describes how often AI Overviews appear. It does not tell us how many of those appearances ended with a click, because neither source reports that figure. The two firms also count differently, so treat 43% and 48% as a range and a direction, not a scoreboard.
Next step. Run ten real searches in one client’s category. Save the answers. Note whether the client appears, and note who appears instead.
Source: TechCrunch, citing Similarweb
3. Publishers Are Considering Opting Out of Google Entirely
What Happened
The Wall Street Journal, reported via Nieman Lab, says major publishers including Reddit, USA Today, Politico, Reuters, and the Economist are evaluating whether to block Google’s crawlers or opt out of search entirely. Some have seen search traffic fall more than 40% between June 2025 and June 2026. Reddit is reportedly reconsidering its $60 million per year deal with Google, and Cloudflare is preparing to block multi-purpose crawlers by default.
Why It Matters
Every AI answer is assembled from something. If a meaningful share of the large publishers that supply that something decide to leave, the answer layer narrows toward whatever sources remain.
Here is my inference, and I will label it as one. If content supply tightens, answers about a client’s category may get thinner and more dependent on a smaller set of sources. Which sources those are would matter more, not less. That cuts both ways for a local business. Less noise, but also less room to be the interesting exception.
There is also a plain business signal in the Cloudflare detail. Blocking by default, at the infrastructure layer, turns a publisher policy meeting into a checkbox. More sites will end up opted out without anyone holding a vote.
The Limit and the Next Step
Where this stops. Blocking is a bet. A publisher that blocks Google gives up the referral traffic that still arrives in exchange for leverage. Not every publisher can afford that trade, and no one named in the reporting has committed to making it.
Next step. Identify the three or four publishers that consistently cover your client’s local market or trade niche. Those are the sources worth watching, because they are the ones an assistant is most likely to surface.
Source: The Wall Street Journal, via Nieman Lab
4. EU Antitrust Regulators Are Quizzing Publishers on Google’s AI Opt-Out Proposal
What Happened
Reuters reports that EU antitrust regulators are seeking publisher feedback on Google’s proposal to let publishers opt out of AI search without affecting their search rankings. The proposal rolled out globally by August 31. That feedback could determine the outcome of an ongoing EU investigation, which could result in another hefty fine if the proposal fails to address competition concerns.
Why It Matters
This is the mechanism story in the round-up, and it is the one most likely to produce something concrete.
The hard part of the publisher complaint was never the opt-out itself. It was the trade. Publishers argued they could not refuse AI use of their content without losing the rankings and traffic they depend on. Google’s proposal tries to split those apart. Refuse the AI use, keep the rankings. Whether that actually holds is exactly what regulators are asking publishers to assess.
If the proposal survives scrutiny, opting out becomes a normal setting on a website rather than a protest. If it does not, the EU has shown before that it is willing to fine.
The Limit and the Next Step
Where this stops. A proposal is not a ruling, and asking for feedback is not the same as finding a problem. The investigation could end with the current plan, with changes, or with a penalty. None of that is decided.
Next step. Watch for the opt-out to appear as an actual control in Google’s publisher tools. When it does, the useful question is whether anyone outside Google can verify that opting out of AI use leaves rankings untouched.
Source: Reuters
5. Yelp Licenses Its Reviews and Reservations to OpenAI
What Happened
Axios reports that Yelp is licensing its reviews, ratings, photos, and business data to OpenAI so that ChatGPT surfaces Yelp content in local recommendations. Yelp Reservations and Waitlist are now integrated directly into ChatGPT.
Why It Matters
Here is where this changes. Four of these five stories are about conflict. This one is about a deal, and the deal tells you what the assistant layer is built from. Licensed human review data.
That should reframe how you think about a client’s review profile. Reviews used to be a conversion asset that lived on a listings page and helped a person decide. They are now also an input into a system that answers the question directly, and the systems with the richest local data are the ones doing the answering.
For local search, the practical implication is that AI assistants are becoming a primary discovery layer, not a novelty. That layer needs businesses to be present, complete, and consistently described across the sources it licenses from.
The Limit and the Next Step
Where this stops. Nothing in this report says how ChatGPT chooses among local businesses, or that a Yelp listing produces a recommendation. Nobody outside OpenAI knows that, and anyone who tells you otherwise is guessing.
Next step. Check one client’s review profile for completeness across the platforms assistants license from, starting with the ones their customers already use.
Source: Axios
The Throughline: Google Becomes the Destination, and Everyone Else Recalculates
Read together, these five stories describe a single adjustment happening on five fronts at once.
The answer surface is expanding fast. The legal challenge to it stalled. The content supply behind it is threatening to walk. Regulators are deciding whether the escape hatch is real. And the assistant layer is quietly being stocked with licensed data from the sources that chose to deal rather than fight.
Here is what I have learned watching this market move. The businesses that show up well in AI answers are not the ones with the cleverest trick. They are the ones whose basic facts are consistent, whose reviews are current, and whose content answers a real question. That was true before AI answers existed, and it is still true now.
What changed is where the answer gets delivered, and how few businesses have checked what it currently says about them.
A Practical Check You Can Run This Week
This takes about thirty minutes, needs no new tools, and works for any client you already serve. Do it for one of them.
- Pick a client with a real local or niche footprint and at least a year of history with you.
- List the five questions their best customer would type before hiring them.
- Run those questions in an AI assistant and in Google search. Save the full answers.
- Note three things: whether the client is named, what the answer says about them, and which competitors are named instead.
- Compare the client’s business details, hours, services, and reviews across the sources the assistant cites.
- Write the gaps down in plain language and rank them by how easy they are to fix.
That list is the beginning of an AI presence audit, and it usually surfaces two or three things worth fixing before anything else. It is a testable way to see how a client is described right now, without guessing at anyone’s algorithms.
FAQ
Do AI Overviews hurt local businesses the same way they hurt publishers?
Not necessarily, and nothing in the reporting above claims they do. Publishers depend on traffic to monetize content. Most local businesses depend on being chosen when someone nearby is ready to buy. Those are different problems. The risk for a local business is being described incorrectly, or being left out of the answer entirely.
Should my client block Google’s AI crawlers?
That is a decision for the business owner. The reporting above describes publishers making that call at a scale and with a leverage most small businesses do not have. For most local businesses, being absent from the answer is a bigger risk than being included in it. If a client does want to explore blocking, Cloudflare’s crawler defaults are the place to start understanding the options.
Can an AI platform be forced to recommend my client?
No, and no one should tell you otherwise. These systems are not required to recommend any particular business, and neither the court ruling nor the EU proposal creates that obligation. What you can control is whether the underlying information about the client is accurate, current, and consistent across the sources those systems draw from.
Does the Yelp deal mean ChatGPT only recommends Yelp businesses?
No. The report says ChatGPT surfaces Yelp content in local recommendations after licensing Yelp reviews and business data. It does not say Yelp is the only source, and it does not describe how businesses are ranked or selected.
What is the single most useful number in this round-up?
The jump in AI Overview appearances, from roughly 15% to around 43% of searches in a year. It is the clearest available measure of how quickly the surface your clients compete on has changed.
One Next Action
Pick one client you already serve and run the check above. You will learn more from a real answer about a real business than from any forecast about where this is heading.
This market changes frequently, and RediForAI monitors those changes weekly so agency owners are not rebuilding their approach from scratch every time a platform shifts. The point is not to react to every headline. The point is to know what the answer says about your client today, and to have something useful to do about it.