UK Announces Proposed Measures in Google Investigation

Megan Kirkwood is a fellow at Tech Policy Press.

The UK’s Competition and Markets Authority (CMA) has announced a significant update to its investigation into designating Google as a Strategic Market Status firm under its new ex-ante competition regulation. The Digital Markets, Competition, and Consumers Act (DMCCA) came into force on January 1, 2025, which gives the competition authority the power to designate and impose rules on Strategic Market Status (SMS) firms in relation to a digital activity or activities. Unlike Europe’s Digital Markets Act (DMA), the UK rules are tailored to the specific firm and digital activity, taking the form of conduct requirements and/or pro-competition interventions.

On January 14, 2025, the CMA announced its first SMS investigation against Google, related to its search and advertising business as a single digital activity. On January 23, the CMA announced the second investigation, this time into Google and Apple's “mobile ecosystems,” identified as a single digital activity. Although the CMA had originally planned to launch a third SMS designation investigation within the six months of 2025, the CMA has since announced that the authority will instead focus on progressing the two ongoing investigations. As a result, no additional SMS investigations will be launched this year.

In its June 24 announcement, the CMA provisionally found that Google meets the legal criteria for designation under the new law, in relation to its general search and search advertising activities. This provisional designation includes Google Search – regardless of how it is accessed – including the information it returns through its underlying infrastructure (including generative AI features), as well as results shown on its search engine results page. It also includes Google Ads, SA360, and AdSense for Search when engaged in search advertising. Google’s Gemini AI assistant, however, will not be explicitly included in the designation, according to the CMA.

While this is not a final decision – which isn’t due until October – this announcement seeks feedback on its current proposed conduct requirements and offers a glimpse into what the enforcement of this law might look like. For the most part, proposed interventions largely mirror the European DMA and are slightly underwhelming. That said, the CMA’s proposal to give website owners and publishers more control over their data could be hugely consequential if it becomes a conduct requirement later this year.

Stages of intervention

The CMA released an in-depth proposal outlining its decision and a detailed roadmap, which identifies priority areas for designing conduct requirements. A blog post by CMA Chief Executive, Sarah Cardell, highlights that early interventions include choice screens, mandating fair and non-discriminatory ranking of search results, data portability, and introducing more control and transparency for publishers over how their content is collected for search used in AI-generated responses.

These are identified as Category 1 measures, which seek to quickly address market imbalances. Likely, if Google is designated as a Strategic Market Status firm, these early conduct requirements would be implemented first. However, a CMA press release also explains that:

The CMA plans to consider a second category of actions to address more complex issues over a longer period (starting in the first half of 2026). These include concerns about the impact of Google’s bargaining position on publishers, its treatment of rival specialized search firms, and concerns about transparency and control in relation to search advertising.

Therefore, it is suggested that more conduct requirements or even pro-competitive interventions might be introduced early next year, depending on the outcome of CMA observations. The CMA also mentions a third category, those not yet required, or where action may be considered later if priority interventions do not sufficiently address the issues. These could include measures relating to consumer control over the use of their data, restrictions on Google’s ability to share data within its ecosystem, or measures on ad load, auctions, and ad prices.

Early conduct requirements

Choice screens

The CMA’s roadmap first introduces the possibility of imposing a search engine “choice screen obligation on key access points, such as Chrome and Android”. They acknowledge that in the UK, some users do already see a choice screen when setting up a new Android device; however, this conduct requirement would legally require it for all users. The roadmap frames this as low-hanging fruit, acknowledging that such a requirement has already been mandated in the EU under the DMA.

Going further, the CMA aims to clarify whether choice screens should be extended to AI assistants. This comes as the current antitrust trial against Google Search by the US Department of Justice revealed details of how Google has paid Samsung “an ‘enormous sum of money’ every month to preinstall Google generative AI app, Gemini, on its phones and devices.”

Data portability

The CMA is considering expanding data portability mandates as a potential conduct requirement. Data portability is the ability of individuals to obtain and reuse their personal data for their own purposes across different services. While currently a legal requirement under the General Data Protection Regulation, there are significant shortcomings that have meant data portability has been ineffective in practice. Under the European DMA, designated gatekeepers have had to introduce more advanced methods of data portability, so that users can more easily port their data. The CMA explicitly states that their proposed “measure would build on similar requirements imposed under the DMA in the EU,” meaning that the data portability tool rolled out in the EU would be legally mandated to be extended to the UK.

Fair ranking and effective complaints for search

The roadmap articulates that in the presentation of search results, they “want to establish a ‘fair ranking principles’ requirement on Google to give users confidence that Google’s ranking and presentation of search results is fair and non-discriminatory.” The CMA found, during its investigation, that “Google may not consistently provide fair search ranking and is able to rapidly (and with limited transparency over when or why) introduce changes to ranking and presentation of results which affect businesses’ ability to reach customers.” Therefore, along with a fair ranking principle and commitments to fair ranking, the CMA has proposed that “businesses should have a route to ensure complaints are properly addressed” and that Google should “maintain an effective complaints process.”

There are similar echoes in the DMA, for example, Article 6(5), bans gatekeepers from self-preferencing their own products and services on their vertically integrated ranking, indexing, and crawling platforms. The DMA also mandates that gatekeepers “shall apply transparent, fair and non-discriminatory conditions to such ranking.”

Transparency, attribution, and choice for publishers

Possibly the most novel of the early interventions proposed, arguably the most urgent, would be to give much more control to publishers over their online published content and how it is collected for search or used in AI-generated responses in AI overviews or Gemini. The CMA makes clear its provisional view that publishers should be able to have their content appear in search results without being automatically opted into data scraping for AI-generated responses.

Bloomberg previously reported that Google uses its search index crawler, Googlebot, to both index websites for search and draw in content for its AI overviews. This forces publishers to either surrender their content to AI overviews, which may damage traffic to their site, or not appear in Google’s Search results, a search engine used by 90% of the world’s population. It has also been widely reported that publishers struggle to block AI crawlers, as the only current solution is implementing robot.txt (a protocol used to control which web crawlers are able to index web pages), which does not consistently work. I have previously highlighted that studies dating from as far back as 2010 have found commercial crawlers, like Google’s, “constantly disobey” robot.txt rules. Additionally, Audrey Hingle and Mallory Knodel wrote for Tech Policy Press that robot.txt cannot express granular preferences, like allowing a website to be indexed to be shown in search results but not scraped for AI training. They write that while new solutions are being proposed, they:

...primarily focus on preference signaling rather than enforcement. Technical tools, standards, and protocols provide mechanisms to express content owners' wishes clearly and machine-readably, but they rely on voluntary compliance from crawlers and AI systems. Robust enforcement and accountability are anticipated to emerge from future policy frameworks and regulatory actions, with policymakers expected to play a crucial role in creating legal backing and consequences for adherence or violations.

The CMA's proposal to introduce a legally binding mechanism to allow publishers to express more granular preferences to Google would be a significant step in the right direction. That being said, the CMA is under immense pressure to balance “the rights of content owners and the need to support innovation that consumers value,” reflecting the government’s pro-growth, pro-industry agenda being imposed on regulators, potentially threatening the CMA’s independence. The roadmap points out that the CMA does not have control over the government’s potential amendments to rules governing Copyright and AI, which have been the topic of much debate and controversy, and remain unresolved.

Longer-term interventions

In early 2026, the CMA proposes the introduction of Category 2 measures, which are potential conduct requirements or pro-competition interventions “on which we think there may be a case for action, but where issues require further consideration.” These include monitoring fairness and transparency in ranking, and could mandate further interventions to prevent self-preferencing in specialized services or even self-preferencing its own AI services (e.g. Gemini AI assistant) over competitor AI services.

Other Category 2 measures will expand on its ongoing monitoring of Google’s use of publisher content for generative AI products. The CMA explicitly writes that the government's decision on copyright and AI will influence this proposed measure, which is worrying given that “the government’s preferred option is to give away the property rights of those who earned them on the promise of growth, growth, growth to the nation,” according to Baroness Beeban Kidron. This could water down the most ambitious proposal in the roadmap.

Finally, under Category 2, the CMA will collect more evidence on advertiser control and transparency. For example, CMA will monitor whether Google is influencing ad auction processes to increase winning bid prices, and may mandate increasing the transparency of Google’s operation of the ad auction process, as well as giving advertisers more data on their ad performance and control over where their ads appear. Additionally, to increase transparency on search display, the CMA may introduce further transparency requirements on how ads are presented in order to “improve consumers’ ability to accurately identify organic links, thereby reducing pressure on advertisers to secure ad slots at the top of the page.”

International developments

The CMA acknowledges that they are not operating in a silo. As well as the references to the European DMA mentioned so far, Cardell’s blog post points to the United States v. Google LLC case, for which remedies are still to be decided. She writes that while the CMA considered “a further set of possible actions (for example, restricting use of default agreements and providing access to underlying search data),” they will instead see the outcome of that case, due in August this year.

The roadmap essentially points out that it is probably easier to see the outcome of that ruling before pursuing the same remedies. The danger here is that if the DOJ does not force Google to share click and query data or search indexes, the likelihood that the CMA would do so is questionable. Therefore, some of the more drastic measures to actually undermine Google’s immense power are being left to other jurisdictions to enforce.

Final thoughts

The CMA’s proposal largely encompasses easy remedies, where Google could essentially extend obligations from the EU to the UK, like providing users the choice to select new defaults or port their data to new services.

However, the more ambitious proposal, around data and AI, is about shifting control. Giving publishers more control over how their data is indexed and used, rather than Google maintaining all of the control and bargaining power, is arguably where Google may push back the hardest.

The next steps will include both third-party stakeholder consultation and engagement, as well as consultation with Google, with the design of this regulation requiring Google to agree to the final measures. How watered down the final conduct requirements will be remains to be seen, but assuming the government doesn’t rule to completely remove copyright protection in favor of AI-scraping, the first conduct requirements could give more bargaining power to the websites and publishers that make the web a vibrant and important resource.

Disclosure: The author submitted evidence to the UK Competition and Markets Authority (CMA) to include data portability as a conduct requirement on behalf of the Coalition for Online Data Empowerment. The author does not receive monetary or non-monetary compensation for this work.