Google Under Pressure in UK: Why Regulators Want New Rules for AI Search

Britain's Competition and Markets Authority has proposed new rules for Google Search, focusing specifically on how websites can control what happens to their content when Google uses it in AI-powered search results. The push comes as news publishers report losing traffic since Google launched AI Overviews — summaries generated by artificial intelligence that sit at the top of search results, changing the way people find information online.

Google responded by saying it supports the regulator's goals of fairness and giving publishers more control. The company also introduced Google-Extended, a new tool that lets websites decide how their content is used to train Google's AI language models.

What the UK Regulator Wants

The authority is asking for three main things: publishers should be able to choose whether their content appears in AI-generated answers, Google should be clearer about how it processes content, and the AI results should properly credit the original sources. The CMA specifically wants news sites and creators to have the option to block their material from being used for AI Overviews.

AI Overviews are Google's biggest change to search since "featured snippets" (those quick answer boxes that appear at the top). Instead of just listing links, AI Overviews pull information from multiple websites to answer your question directly. This can be helpful for users—fewer clicks needed—but it worries publishers because fewer clicks means less traffic to their sites, and traffic is how they make money from advertising.

The regulatory push goes deeper than just opt-out buttons. The CMA wants granular controls—meaning publishers could set different rules for different types of AI use while still staying visible in regular search results. It's trying to find middle ground between making search more useful and protecting the people who create the content.

The Technical Problem

Here's where it gets complicated. Google has a tool called Google-Extended that controls how content trains AI models, but that's only part of the picture. When Google creates summaries for AI Overviews, it uses a different technical process than when it's training those models. Both processes involve your content, but in different ways, so they need different control systems.

Historically, publishers have accepted that Google's automated crawlers read their websites—that's how Google Search works. But AI processing is different: it's not just finding your content, it's using that content to create a summary that competes with your original article. That shift means Google needs to build new ways for publishers to control access to their material.

Right now, websites use simple on-off switches like robots.txt files (a basic instruction file that tells crawlers what they can and can't index). That level of control isn't detailed enough for what the CMA wants. Building more precise permission systems while keeping search fast and relevant at Google's scale—billions of pages processed daily—is a genuine engineering challenge.

Why This Matters for News Sites

The traffic drop that news publishers are seeing isn't just about one new feature. It points to a real problem: when search results answer your question directly, you have less reason to click through to the original article. Since most news websites make money from advertising tied to page views, fewer visits means less revenue.

The broader context here is that AI-powered search is changing how people get information much faster than previous shifts. When the internet moved from computers to phones, that took years to fully play out. This AI change is happening in months. Publishers are struggling to adapt their business models while the rules of the game are still being written.

The UK regulator is trying to act before these changes become permanent—before the market gets locked into a new pattern that favors search engines over publishers. If similar rules appear in other countries, it could reshape how AI companies handle content worldwide, though companies might face different requirements in different places.

The Bigger Picture

This debate echoes earlier conflicts over how dominant tech platforms reshape media distribution. Europe's Digital Markets Act sets a precedent for imposing specific technical requirements on large platforms. The UK is taking a different approach through competition law rather than content rules, but the end goal is similar: more fairness in how the system works.

The UK's investigation into AI also overlaps with inquiries into data practices and market power. For companies and publishers trying to navigate this landscape, that creates uncertainty. Nobody quite knows what the final rules will look like.

One notable factor: the UK sits in an unusual position as both a major English-language market and an independent regulator outside the EU. How Britain handles this could set precedent for other countries and influence whether tech companies make different choices in different regions.

What Happens Next

Google has publicly said it agrees with the CMA's goals, which suggests the company is willing to work with regulators. But Google also highlights the technical difficulty of what's being asked. The consultation process gives companies and publishers time to have their say before final rules are written, which might lead to adjustments as people point out what's feasible.

News publishers generally want more control over AI content usage, though they're skeptical of voluntary agreements. The core tension—balancing how much content AI needs to train on against making sure creators get compensated—remains unsolved. Different groups are proposing various approaches, from licensing fees to payment models.

The outcome will likely influence how the European Union, which has broad powers under its AI Act, approaches similar problems. As AI becomes more central to how people access information, companies will need to follow multiple rulebooks in multiple regions.

Looking Ahead

The deeper question underneath all this is how to balance AI's benefits with fairness to the people whose work trains these systems. The CMA's consultation is establishing a principle: regulators should have a say in how AI products are designed when they affect how information reaches the public.

The stakes are substantial for both sides. Publishers might get relief from traffic losses, but they'll need to decide carefully whether blocking AI or cooperating with it serves them better. Too strict, and search engines might de-rank their sites. Too permissive, and they accelerate their own decline. There's no obvious right answer yet.

How this gets resolved in Britain will likely affect similar conflicts across industries and countries for years to come. The decisions made now will shape what it looks like when AI systems that rely on third-party content meet regulations designed to protect creators.