AI news use triples as publishers adapt to answer engines

Use of generative AI to access news has tripled in three years across six countries tracked by the Reuters Institute for the Study of Journalism, amid industry concerns that readers may receive publishers’ reporting without visiting their sites. The share rose from 3% in 2024 to 6% in 2025 and 9% in 2026, while 8% of respondents across a broader nine-country survey said they used AI for news.

The institute’s Generative AI and News Report 2026, published Oct. 29, found that asking follow-up questions about an event was the leading news-related use among AI users, at 49%. Seeking the latest updates followed at 46%. The findings reflect a shift from search results built around links towards systems that synthesise material from multiple sources into direct answers.

The survey, conducted online by YouGov between Aug. 14 and Sept. 28, covered about 2,000 respondents in each of Argentina, Australia, Denmark, Finland, France, Japan, Norway, the UK and the US. Samples were weighted by age, gender, region, education and political leaning.

AI remains a secondary route to news rather than a replacement for established channels. In the six markets with comparable data, 51% used dedicated AI systems weekly, up from 34% in 2025, but generative AI was still used for news less often than search engines, social media, video-sharing platforms and direct visits to news sites and apps. In the US, 12% had used generative AI for news in the preceding week.

The report also found broad exposure to AI through existing services: 57% across the nine countries had seen or engaged with AI answers, summaries or assistants in search during the previous week, and 49% had encountered an AI feature on a news site or app. AI-content labels were the most noticed news-site feature, at 24%, followed by bullet-point summaries at about 20%.

That distribution shift poses a commercial problem for publishers. When an answer engine incorporates journalistic material without clearly identifying its origin, the publisher may lose traffic, audience contact and revenue. Across 27 markets in the Reuters Institute’s Digital News Report 2026, only 4% said they always or often clicked from AI to underlying news sources, compared with 19% from search and 17% from social media. Globally, weekly chatbot use for news rose from 7% to 10%, though only 1% described AI as their main source of news.

News organisations are consequently experimenting with answer-engine optimisation, or AEO, and generative-engine optimisation, or GEO, to increase the chances that AI systems recognise and select their work. The institute defines AEO as methods to help content providers gain visibility in chatbots and other AI-driven interfaces. Its January publisher survey found a net priority score of +61 for navigating distribution through platforms including ChatGPT, Gemini and Perplexity, while conventional Google SEO scored -25.

Those tactics are not presented as a complete answer. The Vietnam analysis that highlighted the new figures said publishers must also sustain direct readership, audience relationships and the value of their reporting. In the Reuters Institute’s earlier survey, publishers expected search traffic to decline by more than 40% over three years and planned to emphasise original investigations, on-the-ground reporting, explanatory journalism and human stories.