UK publishers press CMA for stricter controls on Google’s AI use to protect journalism

UK news organisations challenge Google’s claims over AI training on their content, urging the CMA to impose stricter controls and fairer compensation amid fears of diminishing journalism economics and reduced originality.

Google is facing renewed pressure from UK news publishers over the way it uses their journalism to train and refine its AI products, with media groups arguing that the company’s own submissions to the Competition and Markets Authority understate the commercial damage. The dispute forms part of the regulator’s wider consultation on conduct requirements for Google under the Digital Markets, Competition and Consumers Act 2024, a process intended to improve terms for publishers and increase transparency around search.

At the centre of the row is Google’s claim that there is “no realistic prospect of harm” to publishers from training and fine-tuning its AI models on their content. News organisations disagree. Guardian Media Group told the CMA that the use of publisher material at every stage of model development shows its value, but also reinforces the risk that AI outputs could become substitutes for the original reporting. The Guardian also warned that more frequent fine-tuning could make models more current and, in turn, weaken emerging markets for retrieval-based licensing deals.

Other publishers say Google should not be allowed to bundle different forms of content use together. DMG Media, which publishes the Daily Mail, Metro and The i Paper, argued that separate crawlers and separate opt-out settings are needed for training, retrieval-augmented generation and fine-tuning. The Financial Times made a similar case for granular controls at both directory and page level, while the News Media Association said weak controls could encourage Google to rely on fine-tuning instead of grounding, reducing the need to negotiate with publishers over payment.

The CMA has said it is not currently minded to impose separate controls for fine-tuning, but that position has not satisfied publishers. The News/Media Alliance, a US trade group, said publishers should be able to opt out of fine-tuning altogether, warning that continuous updates based on journalistic content could make Google’s AI responses more complete while keeping users inside its own ecosystem. The group also argued that if Google values the content enough to improve text generation and ranking, then a fair exchange should follow.

Publishers are also pushing the regulator not to delay action on compensation. The CMA said in January that it would wait 12 months to assess the impact of its initial steps before deciding whether to go further on fair and reasonable terms for content. DMG Media and the News Media Association want that timetable accelerated, saying the imbalance in bargaining power between Google and news businesses is already harming the economics of journalism. The CMA has also proposed that Google should make AI Overviews and AI Mode more clearly attribute publisher content, but the NMA wants links that are prominent enough to drive readers back to the original source.

The publishers’ submissions also revisit Google’s site reputation abuse policy, introduced in 2024, which penalised some sites for hosting paid third-party content. DMG Media said the policy had wiped out a revenue stream overnight, while one anonymised publisher told the CMA it may be costing more than £1m a year. Publishers are now asking for stronger notice periods before ranking changes, clearer explanations of major algorithm updates, and a complaints mechanism with teeth rather than what they see as a slow and uncertain after-the-fact review.

Source Reference Map

Inspired by headline at: ^[1]

Sources by paragraph:

Source: Noah Wire Services

Noah Fact Check Pro

The draft above was created using the information available at the time the story first
emerged. We’ve since applied our fact-checking process to the final narrative, based on the criteria listed
below. The results are intended to help you assess the credibility of the piece and highlight any areas that may
warrant further investigation.

Freshness check

Score:
5

Notes:
The article references events from January 2026, with the latest update in April 2026. The earliest known publication date of similar content is January 28, 2026, from The Guardian. ([theguardian.com](https://www.theguardian.com/media/2026/jan/28/uk-media-groups-should-be-allowed-opt-out-of-google-ai-overviews-cma-proposals?utm_source=openai)) The Press Gazette article appears to be a derivative of this source, summarising and aggregating content from The Guardian and other sources. This raises concerns about originality and freshness. The Press Gazette article includes updated data but recycles older material, which is a concern. Given these factors, the freshness score is reduced.

Quotes check

Score:
4

Notes:
The article includes direct quotes from various sources. However, these quotes cannot be independently verified, as they do not appear in the provided search results. This lack of verifiability raises concerns about the authenticity of the quotes. Without access to the original sources, it’s challenging to confirm the accuracy of these statements. Therefore, the quotes score is low.

Source reliability

Score:
4

Notes:
The Press Gazette is a niche publication, which may limit its reach and influence. The article appears to be summarising and aggregating content from other sources, including The Guardian and government press releases. This raises concerns about the independence and originality of the content. The reliance on a single source for verification further diminishes the reliability of the information presented. Given these factors, the source reliability score is low.

Plausibility check

Score:
6

Notes:
The article discusses ongoing regulatory actions by the Competition and Markets Authority (CMA) concerning Google’s use of publisher content in AI Overviews. While the claims align with known industry trends, the lack of independent verification and reliance on potentially recycled content raises questions about the accuracy of the information. The absence of supporting details from other reputable outlets further diminishes the plausibility of the claims. Therefore, the plausibility score is moderate.

Overall assessment

Verdict (FAIL, OPEN, PASS): FAIL

Confidence (LOW, MEDIUM, HIGH): HIGH

Summary:
The article exhibits significant issues with freshness, originality, source reliability, and verification independence. It heavily relies on summarising and aggregating content from other sources, including paywalled publications, without providing independent verification or original reporting. The lack of independently verifiable quotes and the reliance on a single source for verification further diminish the credibility of the article. Given these concerns, the overall assessment is a FAIL with high confidence.