|
Getting your Trinity Audio player ready...
|
Google Faces Fresh UK Pressure to Add AI Assistants to Its Android and Chrome Choice Screens
Britain’s competition regulator is pushing ahead with a plan that could force Google to let UK users pick a rival AI assistant, not just a rival search engine, the moment they set up a new Android phone or open Chrome for the first time. With a final decision expected imminently, the question of whether AI chatbots belong on the same choice screen as traditional search engines has become one of the thorniest sticking points in Britain’s broader crackdown on Google’s dominance.
The push traces back to October 2025, when the Competition and Markets Authority formally designated Google as holding what the UK calls strategic market status in search, a label reserved for firms with an unusually powerful grip on a particular digital market. The basis for that designation wasn’t subtle: Google handles more than 90 percent of search queries carried out in Britain. Once a company receives that designation under the UK’s Digital Markets, Competition and Consumers regime, the CMA gains the power to impose a standing set of ongoing obligations, known as conduct requirements, rather than having to fight each grievance through the courts case by case, the more traditional approach the US has taken with its own antitrust action against Google.
The CMA published its first detailed proposals in January, opening them up for public consultation, and by June had moved from proposal to actual enforcement. The requirements now in force cover several fronts at once: Google has to demonstrate its search rankings, including results generated through AI Overviews and AI Mode, follow fair and non-discriminatory criteria; publishers must be given the ability to opt their content out of training Google’s AI models; and, most contentiously, choice screens letting users pick a different default search service have to become a legal requirement on both Android devices and the Chrome browser, rather than the patchier, Android-only setup that currently exists and gets reassessed just once a year.
It’s that choice screen requirement where things get genuinely complicated, because the CMA has been actively weighing whether AI assistants themselves, not just traditional search engines, should appear as selectable options on that same screen. The idea sounds straightforward on paper: if a UK user setting up their phone can already choose between Google Search and a rival like DuckDuckGo or Bing, why shouldn’t they also be offered a choice between Google’s own Gemini assistant and competitors like ChatGPT or Anthropic’s Claude at that same moment? In practice, the CMA has actually kept Gemini itself out of the underlying strategic market status designation for now, citing limited evidence that consumers are currently using it as a straight substitute for general web search, a distinction that’s made the choice screen question feel inconsistent to several parties weighing in during the consultation.
Publisher groups have been especially vocal in pushing back. The News Media Association told the CMA it had serious concerns about folding AI assistants into the choice screen, warning that because AI assistant interfaces already generate very few referral clicks back to original content compared to traditional search, adding them to the screen could accelerate traffic diversion toward AI competitors that simply ignore publisher opt-out signals, like the long-standing robots.txt standard, undermining the very protections the CMA’s other conduct requirements are meant to give publishers. The Professional Publishers Association raised similar objections, and the Guardian’s parent company argued it would be inconsistent to treat AI assistants as equivalent to search engines for choice screen purposes while simultaneously excluding Gemini from the core designation that triggered these rules in the first place. One anonymous respondent to the consultation suggested a middle path, proposing entirely separate choice screens for traditional search engines and AI assistants, pointing out that roughly a third of all searches are navigational, meaning people are simply typing a brand name rather than genuinely comparing information sources, a use case where an AI chatbot substitute doesn’t really apply.
Google, for its part, has pushed back on the mechanics of the choice screen idea more broadly rather than specifically fighting the AI assistant question. The company has argued that repeatedly interrupting users with pop-up choice screens, on top of the one shown during initial device setup, would simply annoy people, and has instead proposed building a permanent, easily accessible switch inside device settings that lets users change their default search service at any time without repeated prompts. Google has also said it’s developing new controls that would let individual websites opt out specifically of generative AI features like AI Overviews, separate from a full opt-out of AI model training entirely, framing these proposals as a way to work constructively with the CMA rather than simply resisting its authority. The scrutiny doesn’t exist in isolation either. A parallel antitrust trial against Google unfolding in the United States has separately surfaced details about the company paying Samsung a substantial monthly sum to preinstall its Gemini app on Samsung phones and other devices, evidence that’s fed directly into arguments about how deeply Google’s AI products are already embedded into the exact access points, phones and browsers, that the CMA’s choice screen rules are meant to open up to competition.
With a final CMA decision on these conduct requirements expected around mid-October, the coming weeks should settle whether Britain becomes the first major market to formally treat AI assistants as direct competitors to search engines deserving equal billing on a phone’s very first setup screen. Whatever the CMA ultimately decides, Google has signaled it doesn’t plan to accept sweeping new obligations quietly, and how this specific dispute over choice screens gets resolved may end up shaping how other regulators around the world, still largely working out how to categorize AI assistants under existing competition law, choose to handle the exact same question themselves.
