Do people trust AI news chatbots because they tried them, or do they only try them because they already trust them?

This explores whether trying AI chatbots for news shifts people's trust over time, or whether trust mostly decides who starts using them. The short answer is that the corpus can't yet show change over time for news specifically. What it does show is a strong link between trust and use, and that link probably runs mostly in the other direction from the one the question assumes. In a 45-market survey, 44% of chatbot news users trust what chatbots tell them, compared with 17% of non-users. Trust predicts chatbot use more closely than it predicts social media use Does trust in AI chatbots drive news-seeking behavior?. The authors read this as trust acting like a gate: people who already trust chatbots choose to try them. That's a different story from trying them and then coming to trust them. Because the survey is a single snapshot, it can't tell those two stories apart.

The broader picture adds a twist. Weekly chatbot news use rose from 7% to 10%, but only 20% of people trust chatbot answers, against 37% for news in general Why do AI chatbots gain news users but lose their trust?. So use is growing faster than trust. One detail suggests people are hedging: 42% of users click through to the original sources. Many people seem to treat the chatbot as a starting point rather than a final authority.

If trust does grow through use, other parts of the corpus suggest the reason may have little to do with accuracy. Focus groups found that ChatGPT earns trust through its conversational feel: it responds to what you say, it's fast, and its answers are neatly formatted. That kind of back-and-forth triggers the social instincts we use with people, and none of it depends on being right conversationality-affords-trust-in-chatbots-because-contingent-interaction-activa. In the same way, chatbots use language that sounds expert, so trust attaches to how an answer sounds rather than whether it's true Does chatbot language style actually shape how much we trust it?. A 199-person study found that people's existing trust in AI, not whether the answer was correct, decided whether they bothered to check a claim. A warm tone made people more likely to agree with wrong answers Does access to web search prevent overreliance on chatbots?. Taken together, these findings point to a risk: repeated use could reinforce itself, building trust that never gets tested.

The pull in the other direction comes from research on novelty. Long-term studies with the chatbot Mitsuku found that the social pull of chatbot relationships fades predictably as the novelty wears off. The researchers also warn that single-session findings can't be stretched to cover months of use Do chatbot relationships lose their appeal as novelty wears off?. Something similar may happen with trust in AI-written text. Right now, readers rate unlabeled AI-assisted messages as highly as human-written ones. The authors predict this default trust will fade as people become more aware of AI writing, but they admit their one-time study can't show that happening Does trust in unlabeled AI messages decline as awareness grows?.

What you might not have expected to learn: the research most relevant to "trust over time" all comes from snapshots, and that gap matters. The evidence leaves two opposing paths open. Conversational warmth could quietly build trust without anyone checking it, or the novelty could wear off and people could become more skeptical. Nobody in this collection has yet followed news users long enough to see which one wins.