One in ten people on Earth now turns to an AI chatbot for news in any given week — and almost none of them click through to the journalism underneath. That is the headline tension running through the Reuters Institute Digital News Report 2026, the most authoritative annual read on global news consumption, published this month by the Reuters Institute for the Study of Journalism at the University of Oxford. The report finds that weekly use of standalone chatbots such as ChatGPT, Google Gemini, and Perplexity for news rose to 10% of audiences, up from 7% a year earlier. It is the clearest evidence yet that conversational AI is hardening into a genuine news interface — and, at the same time, quietly starving the publishers who supply the underlying reporting of the traffic they depend on.
What the data shows
The 10% figure, drawn from a survey spanning dozens of markets and roughly 100,000 respondents, represents a significant proportional jump in a single year. But the report is careful to frame it as a drift rather than a stampede. "The use of AI chatbots for news is growing quickly but not as quickly as AI use for other purposes," the authors write in the executive summary. Adoption is heavily skewed toward the young and the highly engaged: 16% of people under 35 report using chatbots for news, and 38% of chatbot news users fall into the report's "news lover" segment, versus 22% of respondents overall. These are, in the report's phrase, "power news consumers" with complex, multi-platform news diets.
Growth is also wildly uneven by geography. South Korea, Greece, and Spain all saw chatbot news use double year over year, with usage reaching 14% in South Korea. Yet several major Western markets — the United States, the United Kingdom, France, and Germany — reported no increase at all. In the UK, trust in answers from AI chatbots sits at just 6%, the lowest of any market surveyed, against global chatbot trust of 20% and overall trust in news of 37%.
Crucially, people are not using these tools merely to read headlines. The single most popular feature, cited by 42% of users, is the ability to ask follow-up questions, followed by compiling stories from multiple outlets into a single answer and translating coverage from other languages.
The 4% problem
If usage is the optimistic half of the story, click-through is the alarming half. Across the 27 markets where the question was asked, just 4% of all respondents say they always or often click through from an AI chatbot answer to the original news source. The comparable figures are 19% for search and 17% for social media. Even when measured only among the chatbot users themselves, the rate is 42% — and the report cautions that "this may be an area where there are differences between what people say they do and their actual behaviour," meaning the real click-through rate is likely lower still.
That gap is what keeps publishers awake. The report situates it within what it calls the "Google Zero" problem: "established news providers are concerned about whether the use of AI chatbots for news will make it even harder to attract audiences to owned news properties (and hence engage and monetise them directly)." When a chatbot synthesizes an answer in its own interface, the user's need is met and the onward journey to a publisher's site — where subscriptions and advertising actually generate revenue — never happens.
For now, the bleeding is small in absolute terms. The report notes that referral traffic from ChatGPT, while rising rapidly, remains "little more than a rounding error," with Google delivering roughly 500 times as many referrals from search alone. But the trajectory is what matters, and it points in one direction.
Why it matters
Two structural shifts collide in these numbers. The first is the normalization of AI as a primary news interface. For a small but fast-growing and influential slice of the audience, the chatbot is no longer a novelty — it is the front door, the place where they ask "what happened today" and "explain this to me." Adding chatbots to social media and video networks pushes the combined share of people getting news from third-party platforms to 56% globally, a tipping point past TV and owned news websites.
The second shift is the decoupling of news consumption from news production. For two decades, search and social at least sent some traffic back to the outlets doing the original reporting. AI chatbots break that loop more completely, absorbing the journalism into a synthesized answer while sending almost no readers — and almost no money — back to the source. The report's one note of nuance is that AI's most engaged users are relatively more likely to click through specifically to verify a story or check its source, suggesting AI may at least reinforce the value of provenance among the wary.
But the macro picture is stark: a growing share of the world is consuming the news through a layer that does not pay for it.
What to watch
The open question for the year ahead is whether the 7-to-10 trajectory holds, accelerates, or plateaus — and whether the Western markets that flatlined this year start to move. Equally important is the licensing and traffic fight now underway between AI companies and publishers: if chatbots become the dominant interface while click-through stays near 4%, the economics of original reporting come under acute strain. Watch, too, for how platforms such as Google fold AI answers more aggressively into search, which the report deliberately excludes from its chatbot figures — meaning the true reach of AI-mediated news is almost certainly larger than 10%. The interface is changing faster than the business model can adapt.
“Established news providers are concerned about whether the use of AI chatbots for news will make it even harder to attract audiences to owned news properties.”— Reuters Institute Digital News Report 2026, University of Oxford