What UK publisher groups are asking the CMA to do
The PPA, Professional Publishers Association, joined by the Independent Media Association and Movement for an Open Web, has formally asked the UK Competition and Markets Authority to keep AI chatbots off the choice screen that Google will be required to show under its new conduct requirements. The PPA said in its filing to the CMA's consultation on Google's proposed conduct requirements that a screen designed to let Android and Chrome users pick a default search provider should not list ChatGPT, Perplexity, or similar AI answer-engines alongside Google, Bing, and other traditional search engines. The filings escalated publicly around August 17, 2026, as trade bodies pressed the CMA before it finalizes the conduct requirements under the Digital Markets, Competition and Consumers Act, or DMCCA.
The consultation sits inside a long-running case: the CMA has spent much of 2026 working out what conduct requirements to impose on Google's search dominance, and the choice screen, a prompt that lets a user set their default search provider, is one of the oldest remedies in the antitrust toolkit for that market. The SCiDA project's analysis of the consultation responses found stakeholders split sharply on scope, and Digiday reported that Google's own room to exempt AI Overviews from parts of the process leaves publishers unable to safely opt out of AI training and answer features without risking their visibility elsewhere in Google's results.
Why publishers say an equal listing would defeat the remedy's purpose
Publishers argue that adding AI chatbots to the choice screen would undercut the very mechanism the remedy is meant to protect, because a chatbot that answers a query directly keeps the user inside its own interface instead of sending them to a publisher's website the way a traditional search results page does. A Media Operator reported that publisher groups frame the risk as a matter of definition as much as competition: a search engine, in the sense the DMCCA choice-screen remedy was designed around, is a service that returns a ranked list of links, not a service that reads those links for the user and states a conclusion.
The practical stakes show up in where clicks land after each design choice, set out below.
| Choice-screen design | Effect on publisher referral traffic | Effect on remedy's original purpose |
|---|---|---|
| AI chatbots excluded | Preserves existing click-through model, no new diversion from this remedy | Remedy stays focused on search-engine competition, as designed |
| AI chatbots included as equal option | Some users choose an answer-engine by default, reducing link clicks over time | Remedy formalizes AI chatbots as search alternatives, a scope change from its original design |
The real fight: whether a chatbot counts as a search engine at all
Underneath the choice-screen dispute sits a harder question nobody built this remedy to answer: whether an AI chatbot is a search engine for antitrust purposes at all. The choice screen was designed in the search-engine era, when the competitive question was which ranked-link provider a device defaults to, and every assumption baked into that remedy, that users compare providers on relevance and speed, that referral clicks are the currency being protected, presumes a category of product that answers by pointing, not by answering directly.
Whichever way the CMA rules carries its own consequence. Excluding chatbots keeps the remedy's original purpose intact but does nothing to slow the underlying shift of user attention toward AI answer-engines that is already happening outside the choice screen, through Google's own AI Overviews and standalone apps. Including them formalizes AI chatbots as an equal alternative to search on the one surface regulators control, which could accelerate exactly the traffic diversion publishers already fear from AI Overviews, just through a different door. Neither outcome resolves the underlying question; it only decides which door the same shift in user behavior walks through first.
Why this matters beyond Google, and beyond the UK
This fight previews a pattern every EU or UK business that depends on search-referral traffic should expect to see recur: a competition remedy built for the search-engine era being asked, mid-process, to classify an AI-native product category it was never designed to cover. The UK's DMCCA case is the first to reach this specific definitional test in public, but it will not be the last, and any operator relying on a search-era remedy to protect its traffic should assume the same classification fight is coming to its own market.
The EU faces an analogous unresolved question inside its own Digital Markets Act gatekeeper framework, where designated gatekeepers' obligations were likewise written around search and platform competition rather than AI answer-engines; the European Commission has not taken a position on how, or whether, AI chatbots fit that framework, but the same definitional gap sits there unaddressed. Publishers and other search-dependent businesses across the EU should read the CMA's eventual ruling as an early signal of how one major regulator resolves the question, not as a settled answer for their own jurisdiction.
Read next: Google's Watermark Opt-Out Tests the EU AI Act | CMA Moves to Pry Open Apple and Google
