Google Gives UK Publishers Option to Opt Out of AI-Powered Search Results

On June 3, 2026, the UK's Competition and Markets Authority imposed conduct requirements on Google that, for the first time anywhere, give publishers the right to prevent their content from appearing in AI-generated search features while remaining visible in traditional organic results [1]. The ruling marks the most concrete regulatory action yet against AI Overviews — the summaries Google places above conventional search links — which publishers blame for draining their traffic, ad revenue, and long-term viability.

The question is whether the remedy matches the problem. Google commands over 90% of UK search queries and roughly £10 billion in annual UK search advertising revenue [2]. Publishers that opt out avoid feeding Google's AI but forfeit exposure to a feature now reaching 2.5 billion monthly active users worldwide [3]. For an industry already watching its Google referral traffic collapse, the opt-out looks less like a lifeline and more like a forced bet on which form of decline is slower.

What the CMA Actually Requires

The conduct requirements stem from the Digital Markets, Competition and Consumers Act 2024, which took effect on January 1, 2025 [4]. The CMA designated Google as having "strategic market status" in general search services in October 2025, finding that the company holds "substantial and entrenched market power" in the sector [4]. That designation triggered an investigation into Google's treatment of publisher content in AI features, culminating in four specific requirements announced on June 3, 2026.

First, publishers gain a site-wide and page-level opt-out from AI Overviews and AI Mode — the company's newer, fully conversational search interface — without any effect on their traditional search rankings [1]. Second, Google must prominently attribute content used in AI-generated results and include clear links to original sources [1]. Third, publishers who allow their content to appear will receive "clear and detailed" engagement metrics showing impressions, page appearances, and geographic distribution [3]. Fourth, Google must allow publishers to opt out of having their content used for AI model fine-tuning, separate from the display opt-out [1].

Google has nine months to implement the full set of changes, though the CMA has said it expects compliance on key provisions "well before" that deadline [1]. Non-compliance carries penalties of up to 10% of Google's global turnover [5].

The Traffic Collapse That Forced the Issue

The regulatory intervention did not occur in a vacuum. By the time the CMA acted, the data on AI Overviews' impact had become impossible to ignore.

Source: Press Gazette / Reuters Institute

Data as of Nov 1, 2025CSV

Global Google search traffic to publishers fell 33% year-on-year by November 2025, according to Chartbeat data tracking more than 2,500 publisher websites. In the United States, where AI Overviews launched first in May 2024, the decline was steeper: 38% [6]. Google Discover referrals fell 21% globally and 29% in the US over the same period [6].

Independent studies corroborate these figures from multiple angles. Pew Research Center found that click-through rates drop by 46.7% when AI summaries appear — users clicked results 8% of the time with AI Overviews versus 15% without them [7]. A separate analysis of 68,000 search queries found that for keywords triggering AI Overviews, the top organic result's click-through rate fell from 7.3% to 1.6%, a 58% reduction after controlling for general trends [8]. As of early 2026, approximately 58% of Google searches result in zero clicks, with AI summaries contributing significantly to that trend [8].

Google has disputed some of these findings. The company questioned Pew's methodology, arguing that the analysis period overlapped with algorithm testing unrelated to AI Overviews [9]. Google has also pointed to internal data showing that AI features "surface relevant links to help people find information they're looking for quickly" and that "people have been visiting a greater diversity of websites for help with more complex questions" [9]. However, no independent study has confirmed that AI Overviews produce a net increase in publisher traffic.

Small Publishers Bear the Heaviest Burden

The aggregate figures mask a stark asymmetry. The decline falls hardest on publishers least equipped to absorb it.

Source: Chartbeat / ALM Corp Analysis

Data as of Dec 1, 2025CSV

Chartbeat data shows small publishers — those with 1,000 to 10,000 daily page views — lost 60% of their search referral traffic over two years. Medium publishers (10,000 to 100,000 daily views) saw a 47% decline. Large publishers (100,000-plus daily views) experienced a 22% drop [10]. The pattern is consistent: AI Overviews are most effective at answering the specific, informational queries that once drove readers to smaller, specialized sites.

Individual cases illustrate the scale. Digital Trends went from 8.5 million Google-referred clicks per month in March 2024 to 264,861 in January 2026, a 97% collapse [7]. The Verge, HowToGeek, and ZDNet each lost more than 85% of their Google-referred traffic over the same period [7]. Charleston Crafted, an independent lifestyle publisher, reported a 70% traffic loss and 65% decline in advertising revenue [10]. Stereogum, a music publication, lost 70% of its ad revenue in 2025 [10]. Business Insider's organic search traffic fell 55% between April 2022 and April 2025, leading the company to cut 21% of its staff [10].

Local news outlets, already operating on thin margins, face particular vulnerability. They depend disproportionately on search referrals for the informational and utility queries — weather, event listings, local government actions — that AI Overviews handle with minimal attribution. Unlike national titles with established brand loyalty and subscription revenue, many local outlets have no alternative traffic source to fall back on.

The Opt-Out Paradox

The CMA's remedy contains an inherent tension. The opt-out is designed to restore publisher agency, but the market conditions that necessitated it make the opt-out functionally unusable for many publishers.

A publisher that opts out of AI Overviews forfeits visibility in a feature serving 2.5 billion users monthly [3]. Google has committed that opting out will not function as a ranking signal and will not affect traditional search placement [3]. But traditional search itself is shrinking as a traffic source. Media leaders surveyed by the Reuters Institute expect search referrals to decline an additional 43% over the next three years, with one-fifth anticipating losses exceeding 75% [6].

Paul Bannister, Chief Revenue Officer of Raptive (an advertising network serving independent publishers), expressed skepticism about the mechanism's structure: "I think if Google actually wanted to [separate AI crawling from search crawling], they could do it by tomorrow. It's easy...they don't do it because it gives them a competitive advantage" [5].

Owen Meredith, CEO of the News Media Association (which represents UK national and regional publishers), raised verification concerns: "We're skeptical about a remedy that relies on Google to separate data for AI Overviews versus search after it has been scraped" [5]. The NMA had pushed for structural separation of Google's AI and search crawlers — a more invasive remedy the CMA did not require.

The CMA also postponed, for at least 12 months, any formal requirement for Google to negotiate fair content licensing payments with publishers [5]. Industry leaders called this delay "disappointing," as it means the opt-out exists without a corresponding mechanism to compensate publishers who stay in.

Could the Opt-Out Entrench Google's Dominance?

Critics argue that the opt-out framework serves Google's strategic interests more than publishers'. The logic runs as follows: by offering a formal opt-out, Google gains regulatory and legal cover — it can point to publisher "choice" as evidence of fair dealing. But because most publishers cannot afford to sacrifice AI-feature visibility, they remain in the system by default, continuing to supply the content that makes AI Overviews useful without receiving licensing payments.

Iacob Gammeltoft from News Media Europe warned that "Google generally doesn't care about publishers exercising their opt-out...and continues to use their content freely" [5]. Sajeeda Merali, CEO of the Professional Publishers Association, noted that "AI Overviews still replace clicks in many contexts" regardless of opt-out provisions [5].

The structural critique extends to the opt-out's design. Publishers must actively choose to withdraw — the default is participation. This "opt-out" rather than "opt-in" framing matters because behavioral economics consistently shows that defaults determine outcomes for the majority of participants. Publishers skeptics expect Google to "delay and obfuscate" implementation, further tilting the playing field [5].

The Regulatory Pressure Timeline

The CMA's action sits within a broader pattern of escalating global scrutiny. Understanding the sequence matters because Google's concessions have consistently followed — not preceded — regulatory threats.

In August 2024, a US federal court ruled that Google violated Section 2 of the Sherman Act by maintaining an illegal monopoly in search [11]. The Department of Justice's revised proposed final judgment, finalized in December 2025, included provisions preventing Google from penalizing publishers who opt out of AI products [11].

The European Commission opened an antitrust investigation into Google's use of publisher content to train AI models, examining whether Google scraped content without appropriate compensation or opt-out mechanisms [12]. Separately, the Commission's Digital Markets Act proceedings in January 2026 addressed whether AI chatbot providers should have access to Google's search data on fair terms [12]. In April 2026, preliminary DMA findings required Google to provide greater data access to search competitors [12].

The CMA's own timeline moved from investigation launch in January 2025 to SMS designation in October 2025 to proposed conduct requirements in January 2026 to final imposition on June 3, 2026 [4]. Each stage followed public complaints from publisher trade groups and parallel actions by US and EU regulators.

UK Opt-Out vs. Mandatory Compensation: A Global Comparison

The UK's approach stands in contrast to frameworks in Australia and Canada, where governments have pursued mandatory compensation rather than voluntary opt-out.

Australia's original 2021 News Media Bargaining Code forced platforms to negotiate payment deals with publishers. When Google refused to renew multiple contracts with Australian publishers in May 2025, the government responded with draft legislation for a "News Bargaining Incentive" — a 2.25% levy on the Australian revenues of Meta, Google, and TikTok, applying from July 1, 2026, unless those platforms strike payment deals [13]. The revised scheme is expected to deliver $144 million to $179 million annually to Australian news organizations [13].

Canada's 2023 Online News Act required platforms to negotiate payment agreements. Google negotiated a $100 million CAD annual payment — significantly less than the $170–300 million CAD proponents expected [14]. Meta responded by withdrawing all news content from its Canadian platforms in October 2023, a ban that remained in place through late 2025 [14].

The mandatory compensation models have produced concrete payments. The UK opt-out model has produced, so far, zero direct revenue for publishers. The CMA deferred the compensation question for at least a year. Australian publishers receive money. Canadian publishers receive money (from Google, not Meta). UK publishers receive a toggle in Google Search Console.

The counterargument is that mandatory compensation models carry their own risks. Meta's complete news withdrawal from Canada demonstrated that platforms may exit rather than pay, leaving publishers worse off. And Australia's scheme has required repeated legislative intervention as platforms find ways to reduce payments. The opt-out model at least preserves publisher visibility in standard search results without triggering platform withdrawal.

The Case For and Against AI Overviews Benefiting Publishers

Google argues that AI features expand content discovery. The company states that AI Overviews and AI Mode help users "explore content they may not have discovered before" [9]. Google has also highlighted that it now labels content from publications users subscribe to inside AI Overviews, with early testing showing "significantly higher click-through rates on labeled results" [9].

Some data supports a narrow version of this claim. AI-referred visitors spend 68% more time on websites than organic search visitors, and ChatGPT users average 583 seconds per session compared to shorter engagement from traditional search [10]. This suggests that when users do click through from AI features, they are more engaged.

However, the volume tells a different story. AI platform referrals account for just 0.15% of global internet traffic, compared to 48.5% from organic search [10]. ChatGPT referrals grew 200% year-on-year but still represent less than 1% of total publisher referrals [10]. Higher per-visit engagement does not compensate for losing more than a third of total visits.

The strongest case for AI Overviews benefiting publishers applies to genuinely obscure content that users would never have found through traditional search. But the data shows the opposite pattern: AI Overviews are most effective at answering straightforward informational queries — precisely the queries that previously generated the highest volume of publisher traffic [7].

What Comes Next

Source: OpenAlex

Data as of Jan 1, 2026CSV

Academic research on AI's impact on publisher traffic has surged, with 2,374 papers published in 2025 alone — reflecting the scale of the problem and the search for solutions [15]. The practical question is whether the CMA's intervention changes the trajectory.

Google's nine-month implementation window means the full opt-out mechanism will not be available until early 2027 at the latest, though a "subset" of UK publishers have begun receiving access to new Search Console controls [3]. The CMA has signaled that compensation requirements — the piece publishers most want — will follow in a subsequent phase.

Meanwhile, publishers are hedging. The Reuters Institute survey found publishers shifting priorities toward YouTube, AI platforms, and TikTok for distribution, while deprioritizing traditional Google SEO [6]. One-third of publishers surveyed said they plan to block Google AI Overviews once effective tools exist [7].

The UK's opt-out is a meaningful regulatory precedent. It establishes that AI search features require publisher consent, that opting out must not carry penalties, and that content attribution in AI results is a regulatory requirement rather than a courtesy. Whether it addresses the economic harm AI Overviews have already caused — or merely formalizes a choice between two forms of decline — depends on what the CMA does next with the compensation question it deferred.

For small and local publishers, the math is stark. They have lost the most traffic, have the fewest alternative revenue sources, and face the least attractive version of both options: stay in and watch AI summaries replace their clicks, or opt out and lose access to the one growth feature in an otherwise shrinking search ecosystem. The CMA has given them a choice. The market has made that choice largely theoretical.