The UK's competition watchdog has proposed forcing its search-engine choice screen onto Google's AI assistants, a move that would graft a decade-old Android remedy onto the chatbot era.

The regulator — almost certainly the Competition and Markets Authority (CMA), the body UK competition reporters shorthand as "the watchdog" — circulated the proposal in a development Arise News reported. The draft would require Google to display a choice screen whenever a user first interacts with one of its AI assistants: Gemini today, plus any successor product Google positions as an AI search gateway.

What would the proposal actually change?

The framing is narrow: an existing remedy, applied to a new product surface. Three mechanics follow:

• Same picker, new entry point. The Android choice screen fires on device setup and after a Chrome update. An AI-assistant variant would fire on the user's first message to Gemini.
• Same legal hook, plausibly. The CMA's preferred UK statute is the Digital Markets, Competition and Consumers Act 2024, which lets it designate "strategic market status" (SMS) firms and impose conduct requirements. Search choice screens already sit on the UK rulebook as precedent.
• A new product-definition fight. Google's likely rebuttal: an AI assistant is not a search engine, and a forced picker would break conversational flow. Whether the CMA accepts that framing decides the proposal's legal fate.

The shift matters because choice screens have shown measurable, if modest, distribution effects since Android began showing the picker under the EU's 2018 antitrust decision against Google. Moving the same mechanism inside an AI assistant would, in principle, let users redirect generative answers to a non-Google backend.

What this means for SEO

The mechanical question is which backend the assistant draws citations from. Gemini today mixes Google Search results with first-party generative output. A choice screen would let users pick the search backbone that feeds those answers — Bing, DuckDuckGo, Brave, or any vendor willing to integrate.

For SEOs, the shift introduces two monitoring jobs. First, which search engines actually get integrated as Gemini options and on what commercial terms. Second, whether Google's AI-citation logs — when they arrive — segment by chosen backend or only by assistant in aggregate. As of this writing, Google has not, to date, published per-source citation data for its AI products.

Tool-signal speculation, separate from the CMA proposal: third-party rank trackers will need to model whether Gemini citations shift toward the user's chosen search engine, or whether Google isolates the assistant's ranking stack from the picker choice. The proposal itself does not address that.

What's confirmed vs. what's inferred

Confirmed: A UK watchdog has proposed extending Google search choice screens to AI assistants, per Arise News. The proposal is not yet a ruling.

Inferred from background context, not as a direct statement in the source: That the watchdog is the CMA, that the likely statutory basis is the DMCC Act 2024, that Gemini is the in-scope product, and that Google's compliance posture will turn on the AI-assistant-as-search-engine definition.

What to monitor next

• A formal CMA consultation document naming the statutory basis and the in-scope products.
• Google's response and whether it formally contests the AI-assistant-as-search framing.
• Filings from Microsoft, OpenAI, Anthropic, and UK publisher associations as interested parties.
• Any first compliance numbers from the existing Android choice screen published in 2026, which will set the baseline for evaluating the AI-assistant variant.

The proposal is a question, not a verdict: should the picker travel with Google's products as they morph from a search box into a chatbot? The CMA's answer, when it lands, will tell every publisher whether 2026 traffic models need a new variable.