Our podcast: Digital News Report 2026. Episode 4: How people are using AI chatbots for news

In this episode we explore findings from our Digital News Report 2026 around the different ways that people are using AI chatbots for news. We explore how use of AI chatbots for news around the world correlates with other forms of news engagement such frequency of access and trust. We look at what people say the do when using chatbots for news such as summarising, digging deeper or just getting headlines, and we explore data to inform the the discussion around 'zero click-throughs' to publisher websites.

The podcast

Spotify | Apple Podcasts

The speakers

Amy Ross Arguedas is a Senior Research Fellow at the Reuters Institute for the Study of Journalism and works on the Digital News Project including as a co-author of the Digital News Report.

Host Mitali Mukherjee is the Director of the Reuters Institute and is a political economy journalist with more than two decades of experience in TV, print and digital journalism.

Transcript

Who's using AI chatbots for news | Trust in news and use of AI chatbots | How people are using AI chatbots for news | AI and 'zero-click-through' search | Lessons for publishers

Who's using AI chatbots for news ↑

Mitali: Can you do a little bit of scene setting, Amy, in terms of who's using AI chatbots for news and what the underlying trend is, if there is a clear trend visible?

Amy: Yeah, sure. So I think the main big story that we see in this year's data is clear growth in the use of AI chatbots for news. We see ten percent of our respondents globally saying that they've used an AI chatbot for news in the past week, and that's up three percentage points from last year. So while it's still a pretty small minority of the population, it is a substantial relative increase compared to what we saw last year. In terms of who's using it, there's also some interesting detail there. We see the highest usage among the youngest age group. Perhaps unsurprisingly, we know that young people tend to be at the forefront of new technology. At the same time, we've seen growth among some of the older demographics. We actually saw more growth among those who are twenty five to forty four years of age.

The other interesting point, aside from the age difference, is that we're seeing different patterns in different parts of the world. Essentially, growth seems to be strongest in certain parts of Asia. So places like South Korea, where fourteen percent say that they're using chatbots for news. It also tends to be higher in parts of Latin America, parts of Africa, and also in some of the southern and eastern European markets. We tend to see less growth in northern and western Europe, as well as the US, where things have been much more stagnant this year. And in the case of the UK, for example, we have the lowest figure globally. Only four percent say that they've used an AI chatbot for news. Essentially, we see growth happening mostly in places where the use of social and video networks is higher for news. And this suggests that people are potentially building on pre-existing platform habits and integrating chatbots into that. The last point I would make is just that what we see is, at this stage, still very much the use of AI in a complementary function. Only one percent say that AI is their main source of news. And so this is still very much tying into pre-existing habits that people are also using alongside AI chatbots.

Mitali: That's really interesting, Amy. And it is doubly so when you look at a market like the US, where all this AI innovation is being rolled out, but you also have quite a high degree of caution in terms of the use of AI chatbots for news. And the other one to take on board is the point that you made quite a few times in your answer, which is growth. So not an explosion, but some growth year over year. And do we know a bit more about what role interest in news itself plays in the use of AI for news? Just picking up on the last point you made about it being complementary, not swapping one out for the other.

Amy: Sure. Yes. It's quite clear that news interest is a very strong predictor of using AI chatbots for news. We see that it's much more concentrated among the demographics that are really strong news lovers. And it really makes sense, right? Because using a chatbot for news requires proactive steps. You have to go into the chatbot, you have to type something in. So it's quite intentional. And this larger proportion of people who are interested in news also helps explain why we see higher use among people on the political extremes, which is another interesting data point. But this is because they also tend to skew more interested in news. So those who are very left wing and very right wing, respectively, tend to use chatbots more than people in the centre.

Trust in news and use of AI chatbots ↑

Mitali: The other one we've been focusing on for some of these alternatives, Amy, whether it's news creators or social media platforms, is the role of trust while navigating the news in these spaces. How does trust in news correlate with the use of AI for news?

Amy: Yes, there is a clear correlation. And actually, just taking a step back, it's pretty interesting to look at the trust data because generally speaking, if we ask everybody, trust in news from AI chatbots tends to be pretty low. But when we zoom in and look specifically at people who are using chatbots for news, then it looks really different. Trust is actually quite high there. So this is telling us that users seem to think that AI is performing reasonably well at these news-related tasks. We also see a strong relationship, or correlation, between chatbot use and trust at the market level. Essentially, countries that tend to have higher levels of trust also tend to have higher levels of adoption. And this correlation is much stronger between use and trust when it comes to AI chatbots than it is in the case of social media, for example, or even search.

How people are using AI chatbots for news ↑

Mitali: I'm going to move to uses for AI for news. And to be clear, when we're talking about use of AI for news, we know that breaking news, for instance, is not surfaced in AI chatbots at this point. Was it clear what it is that people want out of this process, or what the different ways are that they use AI for news?

Amy: Yeah, we see a broad range of uses, really. I mean, there is a clear frontrunner in terms of the most commonly mentioned use. We see that around four in ten say that they're asking chatbots follow-up questions about news. At the same time, we see a variety of other uses that were mentioned by around a third of respondents. There's a portion of people who are simply asking chatbots for the latest news. There are similar proportions using it to simplify the news, so helping make it easier to understand, and also summarising the news. And then we also see around a third who report using it to help them evaluate news sources. So, for example, asking a chatbot to assess the reliability of a news source, this type of application. So this highlights quite a broad range of uses that go far beyond what's possible on some of the other platforms that we've been talking about for some time.

Mitali: And it reflects the point you were making about news super-users turning to these AI chatbots and taking a far deeper approach into news, evaluating news sources and things like that. You were speaking earlier, Amy, about how this was quite different in terms of different parts of the world. So Asia, probably a bit more culturally experimental with technology, and other use trends as well, ranging quite differently across the world.

Amy: Yeah. I mean, we see some similarities and some differences. I think one of the similarities is about this frontrunner use, which is the follow-up questions. And we see this ranking first in thirty three of the forty five markets where we asked the question. So there's broad consensus that that's in first place. But again, then we do see some differences. And as you were mentioning, in some of the Asian markets, so for example in Taiwan and South Korea, where people are a lot more used to getting news from aggregators, simply asking chatbots for the latest news is the most commonly mentioned use. Then we see in other places, like Canada and the UK, for example, where summarisation comes at the top. In other places, like Austria, we see that making the news easier to understand ranks in first place. And then the other interesting thing has to do with this evaluative use that I mentioned before. Using AI to evaluate news sources tends to rank much more highly, or somewhat more highly, in markets like Hong Kong and Turkey. So places that tend to have lower press freedom scores or lower levels of trust in news. And then the other similarity that we see across markets is that there are some of these other uses, like asking chatbots to change the format, for example, from text into video. Those didn't rank highly in any of the markets. So there's also consensus about those not being as popular.

AI and 'zero-click-through' search ↑

Mitali: One of the things I should have clarified at the start was that when we're talking about using AI to access the news, we are talking specifically about chatbots. So this is not about the AI search mode that you may see on some of your browsers when you open it on Google, for instance. This is specifically the ChatGPT chatbot or Gemini from Google. And keeping that in mind, Amy, there have been quite a few conversations and discussions within the news industry about this zero-click-through concern and how AI is going to change this path of discovery and actually link back to the news websites. It's really hard because we're looking at it from the other end of the telescope, which is what audiences think. We don't have access to publisher data, but what about that? The AI overview space, the click-through question.

Amy: Yes, that's a great point. And yes, of course, it's causing a lot of anxieties about just another platform eating into referral traffic, especially as people are using chatbots to get more detailed, more personalised answers. So in this year's survey, we were trying to get at this question within the limitations. And again, this is a survey, so it's very much self-reported behaviour. People might not even remember very well what they're doing. Sometimes they might report things differently from what they're doing in practice because they think it's socially desirable. So there are limitations to asking people about what they do. But it does give us a starting point. And so what we did was we asked people how often they clicked through from AI chatbots. And then we also asked this in terms of clicking through from search engines and from social media, because that would give us a point of comparison between different platforms. So we asked this across twenty seven different markets. And what we see is that just around four percent of respondents overall say that they always or often click through to underlying news sources from AI chatbots. So this four percent is quite a bit lower than the nineteen percent we see from search and the seventeen percent we see from social media. But again, this is largely being driven just by lower use of AI more generally, because search and social media are used much more generally in the population. So again, this lower percentage is driven in large part by the fact that chatbot use is still largely a niche behaviour among this very engaged group of users.

Lessons for publishers ↑

Mitali: Along with the question for the news industry, I mean, how this impacts the work that they do. It's sort of: what can we bring to the audience that is value-additive by the use of AI? What can publishers take away from this in terms of audiences and their relationship with AI for news specifically?

Amy: Yeah. I mean, I think that there are several lessons here, some about the big picture, which is about this very uneven uptake of AI that we see at the moment. So it's uneven in terms of geographies, but it's also uneven in terms of the kinds of people who are using chatbots. And I think that the fact that early adoption is concentrated among more engaged audiences tells us that some of these early effects are going to be felt potentially stronger among some of the people who are more closely connected to news. It also might add to the challenges around reaching young people, who we know have been somewhat elusive for traditional publishers. Now, in terms of the actual uses that people are giving chatbots, it's clear that AI is not just another route to headlines, even though that is how some people are using it. But in practice, people are using it for a large variety of things, as I mentioned, to ask questions, to summarise, to evaluate information. And this is pointing to a much more expansive role that combines access, but also interpretation. And I think, for publishers, it's clear that some of these uses align quite well with areas where they can respond directly.

And there are some potential lessons here. We see an appetite for simpler, more easily digestible news. We also see people wanting summarising, wanting news consumption to be more efficient. And there's also even some lessons here in terms of the follow-up questions. There could be opportunities here for publishers trying to better anticipate audience needs through explainers or, you know, you see some publishers experimenting with opening up questions and then trying to answer them. So I think that there are certainly some lessons here. At the same time, there are some of these uses that are not as easily addressed. And I think part of the appeal of chatbots really does have to do with their ability to provide these really personalised responses, very low-effort responses at scale. And this is something that's going to be a lot harder for publishers to match. I think that there's clearly some important strategic questions here about where publishers should invest their time and their effort. And it's likely that these efforts to replicate more generic AI functionalities might be outpaced by some of the larger platform companies. So I think here, publishers may be better served by focusing on ways that they can enhance their specific journalism in a way that's distinctive and genuinely valuable for audiences.

Mitali: Amy, thank you very much for joining our podcast today.

Amy: Thank you so much for having me.