More People Read News via AI — Few Trust or Click

Ten percent of people worldwide now use AI chatbots to get their news at least once a week, up from 7% a year ago, according to the Reuters Institute's Digital News Report 2026. The jump is real but the habit is still shallow: only 1% name a chatbot as their main news source, and the growth is concentrated in Asia, Africa, Latin America and parts of Southern and Eastern Europe. Among 18-to-24-year-olds the weekly figure climbs to 17%, against just 5% in the oldest age group — the usual pattern of a young technology spreading from the young outward.

The most consequential number isn't usage — it's what happens to the source. Across 27 markets surveyed, just 4% of people say they always or often click through to the original article after a chatbot answers them. Compare that with 19% for search engines and 17% for social media. In other words, the AI hands you the packaged answer and most people stop there. For anyone who produces the news those chatbots are summarizing, that's the quiet problem: the work still gets read, but the reader never arrives at the door.

Trust splits sharply between users and everyone else

Trust in AI-generated news is low overall — about 20% of the general population — but it nearly doubles to 44% among people who actually use chatbots for news, higher even than the 37% baseline trust in traditional news. That gap is worth sitting with: the more someone relies on a chatbot, the more they trust it, regardless of whether its summaries are accurate. People reach for these tools for concrete reasons, the report found — 42% ask follow-up questions, 35% check current news, 34% want summaries, 33% use it to check how reliable a source is, and 30% to simplify the news. And 39% say AI is simply faster than other ways of getting informed.

Why this matters for you even if you never ask a bot for headlines: the way a whole population learns what's going on is shifting toward a layer that rarely shows its work. A chatbot blends sources, drops the links, and presents one smooth answer — and the data shows most people won't go check it. My honest take after using these tools daily: they're genuinely good for a quick orientation, but treating a single AI summary as 'the news' means trusting an editor you can't see, who never tells you what got left out.