Google's New AI Search Features Draw Regulatory Scrutiny Over Publisher Rights

Google is facing investigations in Britain and Europe over how it uses publisher content in new AI-powered search features. Regulators are asking a straightforward question: should publishers have a choice about whether their articles appear in Google's AI summaries, even if they want to remain in regular search results?

Right now, Google's approach is all-or-nothing. If a publisher doesn't want their content used to train or power AI search features, they can't simply opt out of just that part—they have to choose between participating in AI search fully or being excluded from Google search altogether. Britain's Competition and Markets Authority says this isn't fair. The European Commission is investigating whether the same bundling of AI features with regular search breaks competition rules.

What Google's AI Search Actually Does

Google has rolled out AI Overviews and AI Mode, features accessible through its Search Generative Experience (available at labs.google.com/search). When you search for something, instead of just getting a list of links, you might get a conversational answer that synthesizes information from multiple websites. Google says it counts traffic from these AI features in the same metrics as regular search.

From a technical standpoint, this creates a lock-in problem. A publisher's content gets used in three ways: showing up in traditional search results, being fed into AI training systems to teach the model how to answer questions, and appearing in AI-generated responses. But publishers can't pick and choose—they can't say "use me in regular search, but not for training the AI" or "show my stuff in AI summaries, but without indexing it traditionally." It's one bundled package or nothing. Internal reporting suggests Google made this decision deliberately, drawing what has been described as a hard line against separating these uses.

The European investigation will also look at YouTube. Google uses videos from YouTube creators to help train and improve its AI systems, but creators haven't been given much control or clarity about how their content is being used this way.

Why This Matters to Publishers

Here's the economic tension at the heart of the regulatory concern: AI search summaries answer questions without sending readers to the original article. A traditional search result says "here are ten websites about quantum computing—go read one." An AI summary says "quantum computing uses quantum bits instead of regular bits. Here's how it works," with maybe a link to deeper information. The user gets their answer without clicking through to a publisher's site.

This is a real change in how the web economy works. Publishers have always lived with search platforms sending them traffic in exchange for content. But when the platform itself becomes the answer, traffic can collapse. A publisher's content helps Google train the system that answers questions without sending people to their site.

The concern isn't just "publishers might lose clicks." It's that Google is using its dominant position in search to force an unfair choice: participate in AI training or disappear from search. Smaller publishers, especially those covering niche topics, often depend on Google search for survival. If Google says "this is how AI search works—take it or lose visibility," publishers have little real choice, even though technically they could refuse.

The broader context here is worth understanding. We've seen similar shifts before. When mobile search took off, publishers had to suddenly optimize for small screens and fast loading, or risk traffic collapse. The stakes were high, but at least publishers understood the rules. With AI search, the rules are still being written, and publishers feel locked out of the decision-making.

What Britain and Europe Want

Britain's proposed solution is straightforward: publishers should be able to opt out of AI features while staying in regular search. They can say "use my article in a search result link, but don't feed it to the AI training system." That way, publishers get choice.

Europe is investigating the same issue and planning to tell Google how to handle it. The Commission is also looking at whether Google should share its search data—information about what people search for and what links Google ranks highly—with competing search engines. That would give rivals a fairer shot at building their own search products.

Both regulators see these AI features as different from regular search in ways that matter for competition. It's not just a tweak to search results. It's a distinct product that uses content in a new way.

The Technical Complexity

Any fix Google implements will require real engineering work. It's one thing to say "let publishers opt out." It's another to actually build systems that can respect those choices throughout the entire AI pipeline. When a publisher opts out of AI features, Google needs to make sure their content doesn't wind up in the training data, doesn't get used for AI responses, and doesn't leak through some side door.

There's also the historical problem: much of Google's current AI training data was scraped before any opt-out mechanism existed. If Google implements granular controls, what happens to content that was already used? The company might need to retrain its AI systems using only content from publishers who agreed, or find other technical solutions.

The European proposal to share search data with competitors adds another layer. Sharing query logs and ranking signals helps rivals understand how Google works, which could make search more competitive. But sharing that data carefully—without revealing user information or giving away all of Google's secret sauce—requires careful engineering and legal frameworks. The European Commission's guidance suggests the regulators understand this won't be simple.

It's worth noting that Google is continuing to expand its AI search capabilities even as these investigations proceed. The company recently added tools to remove non-consensual explicit images from search results, showing the platform is still evolving. Any final regulatory fix will need to work alongside whatever Google builds next.

What This Means Going Forward

These investigations matter beyond just Google and publishers. Technology companies are watching to see how regulators will handle AI features in dominant platforms. If Britain and Europe force Google to give publishers real choice and control, that sets a template. Other companies building AI-powered products will face similar expectations: clear user consent, opt-out options, and some form of compensation or control when your content powers AI systems.

The outcomes will also test whether competition authorities can keep pace with how technology companies evolve. AI moves fast. Regulations move slower. Getting this wrong—being either too rigid or too permissive—could either stifle innovation or leave publishers and creators vulnerable. The challenge is finding a middle ground where Google can build compelling AI products while creators retain meaningful control over how their work is used.