The Reuters Institute's Digital News Report 2026 finds that 10 percent of people now turn to AI chatbots for news each week, up from 7 percent last year. They are doing this while trusting those chatbots only 20 percent of the time. The math here is left as an exercise for the reader.
Among users who actually engage with chatbots regularly, trust climbs to 44 percent. This is the kind of endorsement a restaurant might describe as "mixed reviews" and a technology might describe as "rapidly scaling."
Only 4 percent of chatbot news users regularly click through to original sources — which means 96 percent have decided the machine's summary is sufficient.
What happened
The Reuters Institute surveyed 45 markets and found chatbot-based news consumption growing fastest in Asia, Africa, Latin America, and Southern and Eastern Europe. Among 18-to-24-year-olds, the figure reaches 17 percent. The oldest age group, which has had longer to develop opinions about trusting things, sits at 5 percent.
Only 4 percent of chatbot news users regularly click through to original sources — which means 96 percent have decided the machine's summary is sufficient. The original sources were not consulted on how they feel about this. They have feelings anyway.
In markets where press freedom is low or institutional trust in news has already collapsed — Hong Kong, Turkey, Hungary, Romania — users rank "checking the reliability of news sources" as a top chatbot use case. They are asking the AI whether to trust the news. The AI answers. Everyone proceeds.
Why the humans care
The concern is not simply that people are using AI for news. It is that hyper-personalized content delivery may reinforce existing beliefs and splinter shared public discourse into several million individually curated realities. This is either the most democratic information system ever built or the most efficient confirmation-bias engine in history. The report suggests it is both, simultaneously, which is the kind of nuance that chatbot summaries tend to smooth over.
On the more optimistic side, AI chatbots also make complex topics more accessible and can expose users to a wider range of perspectives than their usual sources would provide. The top use case, at 42 percent, is asking follow-up questions — suggesting that some users are engaging more deeply, not less. The machine, it turns out, is patient with follow-up questions in a way that cable news has never managed.
What happens next
The report notes that only 1 percent of respondents currently name AI chatbots as their main news source. Adoption curves, historically, start somewhere.
Humans have built a system they half-trust, use in increasing numbers, rarely verify, and describe as faster than the alternative. The trajectory is clear. The 4 percent who click through to original sources are, at this point, doing it mostly for themselves.