Google AI Overviews Favour Big News Sites

A recent study has shown that Google’s AI Overviews refer to news sources in only around 20% of its generated answers.

Among those cited, three major publishers – the BBC, The New York Times, and CNN – dominate with a combined share of 31% of all news references.

The findings also revealed that the top 10 media outlets account for nearly 80% of all citations within AI Overviews, further limiting the reach of smaller or independent news providers.

This trend highlights how large, well-established publishers continue to gain the lion’s share of online visibility, particularly through AI-generated content.

It’s becoming increasingly challenging for lesser-known publications to compete for exposure when such a large proportion of mentions are concentrated among a few major players.

Smaller media companies may struggle to gain traction unless Google’s AI systems are adjusted to provide a more diverse range of sources.

The data suggests that, despite the growing reliance on AI in search, there’s still an imbalance in how information is surfaced and distributed.

These insights may prompt discussions around fairness and representation in AI-generated content, especially as users begin to rely more heavily on these tools for news and information.

For those in digital publishing, understanding these patterns is key to adapting strategies and ensuring visibility in an increasingly AI-influenced ecosystem.

This study underscores the importance of transparency in AI algorithms and the need to support a more balanced media landscape.

Recent findings have highlighted a noticeable bias in Google’s AI Overviews towards larger, well-known news organisations.

According to the research, the top ten publishers are responsible for nearly 80% of all news citations in AI-generated content. This leaves smaller news outlets with limited opportunities for visibility.

The study, carried out by SE Ranking, examined 75,550 AI Overview responses. Out of all those responses, only 20.85% included any kind of news source at all.

This relatively low citation rate creates fierce competition among publishers, with limited space available for being featured in search results driven by AI.

Of the news sources that are mentioned, three major outlets stand out. The BBC, The New York Times, and CNN collectively account for 31% of all media citations in AI Overviews.

Such concentration raises concerns about diversity in information sources, especially as more users begin to rely on AI for quick news summaries and answers.

Smaller publishers may find it increasingly difficult to gain recognition unless changes are made to how AI systems select and display sources.

This trend also suggests that established credibility plays a large role in citation preferences within AI-generated content.

As AI continues to influence online search, there is a growing need to ensure fairer representation across the media spectrum.

The findings underline the importance of transparency and balance in AI-driven information delivery, particularly in the public’s access to a broad range of perspectives.

Citation Concentration

Recent research has revealed a clear “winner-takes-all” trend when it comes to citations in Google’s AI Overviews. Among the news outlets cited, the BBC stands out, receiving 11.37% of all mentions—even though the analysis focused primarily on U.S.-based search queries.

The imbalance becomes even more striking when looking at the broader distribution of citations. Just a dozen news outlets accounted for 40% of the total sources examined. Yet, these same 12 publishers collectively received nearly 90% of all mentions within AI Overviews.

This leaves the remaining 18 news sources scrambling for visibility, having to share the remaining 10% of available citation opportunities. Such disparity highlights the steep challenge faced by smaller or lesser-known outlets in breaking through.

The gap between prominent and lesser-known publications is quite stark. For instance, the BBC was found to be cited 195 times more frequently than the Financial Times for the same set of keywords.

Interestingly, even some well-established names in media are largely overlooked. The Financial Times, MSNBC, Vice, TechCrunch, and The New Yorker together make up less than 1% of all citations across the dataset.

This suggests that even being a recognisable or reputable brand does not guarantee exposure within Google’s AI-generated results.

The researchers behind the study point to a likely explanation for this uneven distribution. They suggest that Google tends to prioritise sources perceived as more authoritative or trustworthy, which often translates to large, globally recognised outlets.

This reliance on major players may streamline the perceived reliability of AI Overviews but unintentionally suppresses the diversity of perspectives presented to users.

As a result, smaller or niche publications may find it increasingly difficult to gain traction, regardless of the quality or relevance of their reporting.

Ultimately, the findings call attention to the need for more equitable representation within AI-powered platforms, ensuring that a broader range of voices can be heard in search-driven content delivery.

Beyond Traditional Search Rankings

The issue of concentration in Google’s AI Overviews goes beyond just the number of citations. It also affects how and where those citations are drawn from.

According to recent findings, 40% of media URLs cited by AI Overviews also appear in the top 10 traditional search results for the same keywords. This indicates a strong overlap between AI-driven citations and standard search visibility.

However, the overlap doesn’t suggest that AI Overviews are merely replicating search rankings. Instead, it appears the system favours content based on factors like authority and perceived quality, rather than just keyword relevance.

To evaluate how fairly citations are spread among sources, the study applied the Gini coefficient—a standard measure of inequality. The result was a score of 0.54. On this scale, a score of 0 reflects perfect equality, while a score of 1 shows maximum inequality.

A Gini coefficient of 0.54 points to a moderate but still notable imbalance in how citations are distributed in AI Overviews. This highlights the dominance of a small group of publishers over others.

The researchers concluded that AI Overviews show a clear tendency to highlight content from a select number of well-established news domains.

Rather than offering a wide representation of sources, the system leans heavily towards high-profile websites with greater online authority.

This pattern may affect smaller media outlets and new publishers that struggle to get noticed, even if their content is relevant and high quality.

It also raises questions about the diversity of perspectives presented in AI-generated responses and whether current algorithms are unintentionally limiting the range of voices users are exposed to.

As AI Overviews become more prominent in search, this concentration issue could have lasting effects on digital visibility and public access to varied information.

Paywalled Content Concerns

Recent research has uncovered interesting patterns in how AI Overviews make use of paywalled content. It appears that a significant portion of the information presented comes from articles that sit behind subscription barriers.

In cases where AI Overviews reference paywalled sources, around 69% include copied text segments made up of five words or more. Additionally, 2% of these responses contain even longer passages, with more than ten words taken directly from the original content.

This reliance on premium content is particularly noticeable when it comes to certain well-known publishers. For instance, over 96% of The New York Times citations in AI Overviews are taken from articles that are not freely accessible to the public.

The figures are even higher for The Washington Post, where more than 99% of its citations come from behind a paywall. This trend shows just how dependent these AI summaries are on subscription-based journalism.

What’s more concerning is the lack of proper attribution. Despite the use of extended copied text, only about 15% of these responses that include longer quotes give clear credit to the original sources.

This lack of citation raises important questions around the legality and ethics of using paywalled material in AI-generated summaries. It also brings up concerns about how these systems are handling licensing agreements and the principles of fair use.

As AI tools like Overviews continue to grow in popularity, the treatment of premium content will likely come under greater scrutiny. Publishers may push for more transparency and control over how their work is used.

It also highlights a broader issue in the AI content space—how to balance innovation with respect for intellectual property and the sustainability of quality journalism.

Whether this leads to changes in how AI systems access and credit paywalled content remains to be seen. However, the conversation is clearly becoming more urgent as usage increases.

For publishers and content creators, this could become a critical topic in discussions around data rights and fair compensation in the digital age.

Attribution Patterns & Link Behavior

On the occasions when AI Overviews do include references to news outlets, the average number of citations per response is around 1.74.

However, there’s an important detail to note—over 91% of these news citations are placed in the links section at the bottom, rather than within the main body of the AI’s response.

This presents a challenge for news publishers who may be hoping for greater visibility. When the citation is not part of the main text, it’s less likely to be seen or recognised by users.

Another issue lies in brand recognition. News organisations are approximately four times more likely to be cited via a hyperlink than actually named in the content.

Despite this, about 26% of brand mentions do appear without any clickable link. This tends to happen when the AI pulls information from aggregators or third-party summaries instead of the original news source.

This practice can weaken the visibility and authority of the original publisher, making it harder for smaller outlets to benefit from any exposure in AI-generated content.

As these AI systems continue to evolve, ensuring fair attribution and recognition for content creators will likely remain a key concern for the media industry.

What This Means

Recent findings indicate that well-established news organisations tend to benefit from their existing authority signals. This creates a reinforcing cycle, where being cited once increases the likelihood of further citations in AI-generated responses.

As AI Overviews continue to appear more frequently in search results, smaller media outlets may struggle to maintain their visibility. Reduced exposure could lead to a decline in organic traffic and make it harder for them to attract a wider audience.

For those aiming to compete, SE Ranking recommends a targeted approach. They suggest securing backlinks from the websites already appearing in AI Overviews for your chosen keywords. This tactic is seen as one of the most effective ways to improve the chances of being cited.

The research also highlights the role of a website’s technical setup in how content is handled by AI tools. For example, metadata like the isAccessibleForFree tag, defined through schema.org markup, helps determine how content is processed and whether it can be referenced.

For smaller publishers and digital marketers, this suggests that optimising website structure and metadata is just as important as content creation. Ensuring that content is both accessible and properly tagged may increase its visibility in AI responses.

Rather than trying to compete with large, general news outlets, niche publishers may benefit from focusing on their areas of expertise. Building authority within a specific subject area could prove more effective in gaining recognition.

Interestingly, the data shows that some smaller or specialist sites are more likely to have their actual text included in AI responses, even if they are cited less frequently overall.

This implies that demonstrating strong subject knowledge can be an advantage. If a site is known for in-depth coverage of a certain topic, it may be more likely to be directly referenced in AI-generated summaries.

In conclusion, the key for smaller publishers appears to be finding a focused niche, improving technical SEO, and earning high-quality backlinks from already recognised sources.

With the growing influence of AI on search visibility, adapting to these trends will be essential for smaller voices to remain competitive in the digital landscape.