Finance — Artificial intelligence is becoming part of the way news is created, summarized, and distributed to media organizations around the world. But new data suggests that audiences are far more hesitant than publishers, revealing a widening gap between newsroom experimentation and reader adoption.
According to the Reuters Institute’s Digital News Report 2026, just 10 percent of people in 45 markets use AI chatbots to read news, up from 7 percent in 2025. Usage varies widely by country, reaching 14 percent in South Korea and Turkey and 13 percent in Brazil, but falling to 4 percent in the United Kingdom and 6 percent in the United States.
The report, based on nearly 100,000 survey responses, highlights global trends. AI is increasingly present in the news ecosystem, but it has yet to become central to how most people access information.
Newsrooms cautiously embrace automation
Although public adoption remains limited, media organizations are expanding their internal use of AI tools.
German and British public broadcasters are starting to use artificial intelligence to generate article summaries. In countries like Sweden, Colombia, and Mexico, newsrooms are experimenting with automated workflows, including tools to help journalists research archives and develop story ideas. in El Economista Mexico uses an AI system to scan archives and suggest angles for tracking reports.
In Nigeria, AI-based transcription and translation tools are being developed to support reporting in local languages, and fact-checking organizations in the Philippines are implementing AI-assisted verification systems.
However, the Reuters Institute emphasizes that most of these applications remain behind the scenes, rather than reporting to audiences on behalf of journalists.
Shift to “AI as infrastructure” rather than authors
The report’s findings are consistent with broader industry patterns. In other words, artificial intelligence is becoming integrated as an infrastructure rather than a replacement for editorial judgment.
News organizations are using AI to accelerate production cycles, summarize long documents, and process large datasets. However, complete reports generated by AI remain limited, in part due to concerns about accuracy and accountability.
An independent investigation reinforces these concerns. A 2025 study by the BBC and the European Broadcasting Union found that AI chatbots made significant errors in the majority of news-related responses, including incorrect sources and factual inaccuracies.
Recent academic reviews suggest that even the most advanced systems can achieve high accuracy for structured questions, but still fail when the queries contain ambiguity, misleading assumptions, or rapidly changing facts.
Accuracy issues: speed and validation
The central tension surrounding AI-generated news is verification, not speed.
AI systems are good at summarizing and pattern recognition, often producing fluent and coherent explanations within seconds. However, multiple studies have shown that they do not consistently verify facts the way human journalists and editorial processes do.
Rather, errors often arise from gaps in search or sources rather than failures in reasoning. This means that if the underlying data is flawed or missing, the system can confidently present incomplete or inaccurate information.
This poses a structural problem for newsrooms. While AI can scale up production, it does not inherently guarantee factual integrity.
Are articles written by AI more engaging?
Early evidence is mixed.
On the other hand, AI-generated summaries can improve accessibility, especially for complex or technical topics, by simplifying language and shortening long reports. This can increase engagement for viewers who prefer quick consumption formats.
However, engagement does not necessarily equate to trust. Even when using digital platforms for discovery, audiences still prefer established news brands for authenticity, according to research cited in the Reuters Institute report.
Research shows that rather than eliminating bias, AI redistributes it and can make bias less visible because it is embedded in the output of the algorithm rather than the human author.
At the same time, some media researchers argue that AI could reduce certain forms of individual editorial bias by applying a more consistent framework across articles. The ultimate impact remains to be determined.
Public skepticism remains strong
Despite increased exposure to AI tools, public skepticism remains high. In several surveys, a majority of respondents say they are uncomfortable with news that is entirely generated by AI, especially political or sensitive topics.
According to a Pew Research survey released in 2026, most Americans believe AI is advancing too quickly, even as the use of chatbots continues to increase in daily tasks.
This tension between growing use and persistent concerns defines the current stage of AI in media.
