The rapid rise of generative AI has become a growing focus for journalism, as publishers and platforms grapple with what it means for how people access and engage with news. Much of the attention has so far centred on how newsrooms can use AI to produce or distribute content more efficiently. But at the same time, a small but growing share of the public is beginning to use these tools directly to get news, as chatbots become more embedded in everyday digital life.

Recent evidence points to a sharp increase in AI use overall. The Generative AI and news report 2025 found that weekly use of generative AI tools across six markets grew from 18% to 34% between 2024 and 2025 (Simon et al. 2025). In this chapter, we examine the emergence of standalone AI chatbots such as ChatGPT and Google Gemini as a source of news, looking at who is using them, what they are used for, and how these behaviours may shape engagement with, and traffic to, original sources.

Listen to a podcast on the chapter

Spotify | Apple Podcasts | Transcript

Who is using AI chatbots for news?

In our data this year, we find evidence of rising weekly use of AI chatbots for news, up from 7% to 10% globally since last year, driven largely by growth in parts of Asia, Africa, and Latin America, as well as Southern and Eastern Europe – markets where platformisation of news is stronger. While this is still a fairly small minority of the population, the figure represents a substantial relative increase and indicates that AI is beginning to play a more meaningful role in news consumption alongside established pathways. However, only 1% say AI is their main source of news, suggesting it currently plays a complementary role for most users.

Uptake of AI chatbots, both in general and for news, continues to be driven by younger people, typically at the forefront of tech adoption, including AI. The proportion of respondents in the youngest age group using chatbots for news (17%) is three times higher than that in the oldest age group (5%), although the most significant growth relative to 2025 was recorded among those 25–34 (up 4pp).

Much of this uptake is concentrated among people already engaged with news. Usage is considerably higher among news lovers (18% among these most intensive news consumers compared to 7% among those who get news just once a day) and is also higher among those with high interest in news (19% among those who are extremely interested relative to 7% among those not very interested). This helps explain why we also see higher use among those on the political extremes, who also tend to skew more interested in news: 16% and 15% among those very left and right wing, respectively.

Another factor shaping the uptake of AI chatbots for news is how much people trust the quality of their outputs. In a context of already low trust in news (37% of people trust most news most of the time), this year's data show that trust in news from AI chatbots among the general population is lower still at just 20% globally. However, a different picture emerges when looking at those who use these tools: 44% of AI chatbot users express trust in news from AI chatbots, compared with just 17% of non-users, which highlights both the extent to which low trust is driven by people who aren’t using the technology and the extent to which users think it performs reasonably well.

When we plot AI chatbot usage against trust at the market level, we see a strong relationship: markets with higher trust in AI chatbots for news also tend to report higher levels of use. What’s more, as we can see in the next charts, the relationship between trust and use is considerably stronger for AI than for social media, as reflected in the higher R². The dots sit much closer to the trendline in the first chart, meaning that levels of trust are more consistently proportional to levels of AI chatbot use than for social media. A similar pattern is evident at the individual level. People who trust AI chatbots are several times more likely to use them for news, whereas the relationship is weaker for social media.

This likely reflects differences in how these platforms are used. Using AI chatbots for news is still an emerging and more deliberate behaviour, meaning trust plays an important role in whether people choose to use them at all. By contrast, social media is used for a wide range of purposes, and news is often encountered incidentally, so people may continue to use it for news even if they have lower levels of trust.

AI chatbot uses and motivations

So, what exactly are people turning to AI chatbots for when it comes to news? Across 45 markets, asking chatbots a follow-up question is the clear front-runner, reported by 42% of users. A second tier of uses is reported by roughly a third of respondents, spanning several distinct types of uses. Around a third use AI to get the latest news (35%). Similar proportions use it to help simplify news consumption, with 34% using chatbots for summarisation and 30% to make news easier to understand. At the same time, around a third (33%) report a more evaluative use, asking chatbots to assess the reliability of a news source. This highlights the variety of different ways AI can be used for news, beyond what is possible on other platforms.

We can identify some similarities and differences across markets by looking at the rank order of these uses. Across most countries, the dominant use of AI chatbots for news is interrogation, with asking follow-up questions ranking first in 33 of the 45 markets where this question was asked, including Brazil (see chart above). However, other uses are relatively more prominent in different contexts. In Asian markets such as Taiwan and South Korea, where many people are accustomed to getting news from aggregators, simply getting the news is the most frequently cited use. In Canada and the UK, summarisation ranks highest, whereas in Austria the most frequently reported use is making news easier to understand, an application that also ranks highly in Germany and Japan.

Meanwhile, using AI to evaluate news sources ranks highly in markets like Hong Kong and Turkey, which score low in terms of press freedom, as well as markets with lower levels of trust in news, including Hungary and Romania. Finally, some uses remain relatively marginal across countries: asking questions about how the news media work, and converting content between formats, do not rank among the top three uses in any market surveyed.

Another way to understand the uptake of AI chatbots for news is by foregrounding not how people use them but why. Globally, the most commonly cited motivation for using AI chatbots is wanting more depth or explanation (42%), highlighting the interactive nature of these tools. This is followed by an efficiency-driven motivation, with 39% saying AI is faster than other ways of getting news. Several motivations, each mentioned by around a third of respondents, point to AI’s role in helping users process and navigate information: 36% use it to summarise complicated stories, 35% value the sense that they can get an answer to any question, another 35% use it to compile stories from different outlets, and 33% to translate news into their preferred language. In contrast, platform-oriented motivations are less widespread, with 24% saying chatbots are the first place they go for most things and 23% saying they simply prefer interacting with a chatbot. 

Clicking-through from AI chatbots to original sources

One of the key concerns among publishers is that growing use of AI chatbots for news will increasingly eat into referral traffic as users get more detailed, personalised answers to their questions within the chatbot environment. In this year’s survey, we tried to get a better handle on this issue by asking those who use standalone AI chatbots for news how often they click through to original news sources. We asked the same of people who use search engines, and also social media, for news to provide points of comparison.

Before turning to the findings, it is important to acknowledge that the data reflect self-reported behaviour, which may under- or overestimate what people do in practice, due to limitations in recall and the tendency to give socially desirable responses. However, the figures give us a starting point to understand differences in click-through intent. It is also worth noting that the rollout of AI-generated summaries in search engines like Google, Bing, and Naver may already be shaping the experiences – and responses – of those using search for news.

This issue underpins industry anxieties about ‘answer engines’, which publishers fear will increasingly satisfy users’ information needs directly within platforms, potentially reducing click-throughs to original sources. Implementation varies across markets, and Google’s AI Overviews,¹ for example, are disabled for certain types of news queries, particularly breaking or developing news, with the Top Stories carousel instead featuring original reporting (see images below).² However, to the extent that search queries go beyond breaking news, it is possible that the integration of AI features may already be shaping click-throughs for search.

Screenshots comparing different Google queries about the Iran war

The data in the two following charts are for the full sample, meaning they include non-users (in light grey), to allow more meaningful comparisons across platforms with very different user bases. We see that, across 27 markets where the question was asked, just 4% of respondents overall say they always or often click through to underlying news sources from AI, compared with 19% from search, and 17% from social media. The large gap between AI and the other platforms reflects the much smaller proportion of people who use AI chatbots for news in the first place (9% across 27 markets), which as we’ve seen, continues to be a niche behaviour among a small, albeit highly engaged, group of users.

Comparing across markets, differences are fairly small when it comes to AI chatbots, with the proportion of people saying they always or often click through to the original source in the lower single digits in nearly all the 27 markets where we asked the question (e.g. just 1% in the UK compared with 3% in Spain and 4% in Argentina). South Korea, where AI use is above-average and skewed towards an especially engaged subset of people, stands out with a relatively higher figure (8% say they always or often click through), though still low in absolute terms.

We see more pronounced differences when looking at social media and search, which tend to have larger user bases for news. In parts of Latin America and Southern Europe, higher proportions say they always or often click through from social media (e.g. 31% in Argentina and 22% in Spain), compared with much lower levels in Northern and Western Europe, including the UK (9%), where social media use for news is lower more generally. Asian markets sit closer to the middle on this measure, with 16% in South Korea saying they always or often click through from social media. Search follows a similar regional pattern, though the differences between countries are somewhat narrower than for social media. This likely reflects the more intentional way people use search to access news, where users are typically seeking out information and therefore more likely to click through in a more consistent way across markets.

These findings are consistent with a broader pattern where platforms tend to play a more limited role in news access in Northern and Western Europe, but are more central to how people do this in Southern Europe, Latin America, and parts of Asia. Differences may also reflect variation in platform environments. In South Korea, for example, search portals such as Naver combine search with news aggregation and other content within a single environment and play a much larger role as gateways to news than search engines in most other markets.

We also asked users of AI chatbots, search, and social media what motivates them to click through to original news sources and stories. Do users simply want to know more? Or are they looking to verify the information? Or learn more about the source?

Our data across 27 markets show that while wanting more detail is the primary motivation for clicking through among users of all three types of platforms, AI chatbot users are less likely to say they click through because they want more detail (51% of those who say they click through), compared to users of search (59%) and social media (60%). This suggests users may be slightly more satisfied with the amount of information provided within chatbot environments than users of search and social media are in the context of those platforms.

However, AI chatbot users are more likely than search or social media users to say they click through because they want to verify the news or find out more about the source, suggesting click-throughs from AI chatbots may be driven more by doubt or curiosity about the accuracy or provenance of the information. The differences between motivations are, thus, much smaller for AI chatbot users than they are for social and search users, where wanting more detail is a much more prominent reason to click through.

Conclusion

While only a small minority of people currently use AI chatbots for news, rapid growth suggests these tools could become a more significant part of news use over time. As with other emerging technologies, early adoption is concentrated among more engaged audiences, meaning initial effects are likely to be felt most among those already closely connected to news, as well as among younger people.

AI is not simply another route to headlines. While some use it to access the latest news, many are using it to interrogate, summarise, and evaluate information, pointing to a more expansive role that combines access with interpretation. Some of these uses align with areas where news organisations can respond directly. The demand for simpler, more digestible news reinforces the value of clearer writing and summarised formats for time-poor audiences, while the appetite for follow-up questions points to opportunities to better anticipate audience needs through formats such as explainers or newsletters.

But other uses are less easily addressed. Part of the appeal of chatbots lies in their ability to provide highly personalised, low-effort responses at scale, something individual publishers are unlikely to match. This raises important strategic questions about where publishers should invest. Efforts to replicate generic AI functionality risk being outpaced by platform companies, suggesting publishers may be better served focusing on ways of enhancing their journalism that are distinctive and genuinely valuable to audiences. Uses focused on evaluating sources or explaining how news works also sit uneasily with journalism’s role, where such assessments may be viewed with scepticism given publishers’ vested interests.

In terms of click-throughs and referrals, our self-reported data suggest AI chatbot users appear somewhat less driven by a need for detail and more by a desire to verify information or understand its source. Though the picture is still emerging, this indicates that AI chatbots may reduce some forms of referral traffic while reinforcing the importance of verification and information provenance, at least among its highly interested user base. The impact of AI is likely to be uneven, varying across markets and user groups, especially as these tools become more embedded in search and other environments. For publishers already struggling with weakening links to audiences, this suggests a need to focus not only on visibility, but on what makes their journalism distinct, including their role as trusted sources in a noisy information environment and deeper reporting that goes beyond what AI can easily replicate.

Footnotes

¹  Per Google’s documentation in March 2026, AI Overviews had been rolled out in all the markets covered by our survey, except France: https://support.google.com/websearch answer/14901683#zippy=%2Chow-to-control-your-data%2Ccountries-and-territories 

² https://www.google.com/intl/en_us/search/howsearchworks/how-news-works/