TLDR: A comprehensive international study involving 22 public broadcasters has found that leading AI chatbots, including ChatGPT, Google Gemini, Microsoft Copilot, and Perplexity AI, provide inaccurate or poorly sourced information in 45% of their responses to news-related queries. The findings highlight significant concerns regarding the reliability of these AI tools for journalistic and factual content.
An extensive international investigation, coordinated by the European Broadcasting Union (EBU) and involving 22 public service media organizations from 18 countries, has revealed a concerning trend: AI chatbots frequently misrepresent news content. The study, released on October 22, 2025, found that nearly half (45%) of all responses generated by leading AI assistants—ChatGPT, Google’s Gemini, Microsoft’s Copilot, and Perplexity AI—contained at least one significant issue when queried about news events.
The research, which evaluated 3,000 AI responses between late May and early June, identified widespread problems with both factual accuracy and sourcing. Approximately 31% of the AI-generated answers suffered from serious sourcing issues, including missing, misleading, or incorrect attributions. Furthermore, a substantial 20% of responses contained outright factual errors, such as incorrect dates, names, or fabricated events. In some instances, AI systems even created plausible-looking but non-existent news article links.
Specific examples of inaccuracies cited in the study include AI chatbots incorrectly identifying Olaf Scholz as the German Chancellor when Friedrich Merz had already taken office, and naming Jens Stoltenberg as NATO Secretary General after Mark Rutte had assumed the role. Google’s Gemini performed notably worse than its counterparts, with significant issues in 76% of its responses, largely attributed to poor sourcing.
The methodology for this unprecedented study was modeled on an earlier BBC experiment from February 2025, which also found significant issues in over half of AI-generated answers, with nearly one-fifth of those citing BBC content introducing factual errors. Journalists from participating organizations, including the BBC (UK) and NPR (US), posed a standardized set of 30 questions to each AI platform, assessing responses based on factual accuracy, quality of sourcing, clear separation of fact from opinion, neutrality, and contextual relevance.
Jean Philip De Tender, Deputy Director General of the EBU, emphasized the systemic nature of these failures. “This research conclusively shows that these failings are not isolated incidents,” he stated. “They are systemic, cross-border, and multilingual, and we believe this endangers public trust. When people don’t know what to trust, they end up trusting nothing at all, and that can deter democratic participation.” Pal Nedregotten, Technology Director at Norway’s NRK, expressed similar concerns, noting that the results did not instill confidence in loosening control over AI companies’ access to their content. Consequently, NRK has decided to permanently block AI companies from scraping its website content.
Despite these significant deficiencies, AI assistants are increasingly being used as a source of information. The Reuters Institute’s Digital News Report 2025 indicates that 7% of online news consumers use AI chatbots for news, a figure that rises to 15% among individuals under 25. Peter Archer, Head of AI at the BBC, acknowledged the potential of AI but stressed the critical need for trustworthiness. “We want these tools to succeed and are open to working with AI companies to deliver for audiences and wider society,” he said, while underscoring that “people must be able to trust what they read, watch and see.”
Also Read:
- Brown University Study Uncovers Widespread Ethical Lapses in AI Mental Health Chatbots
- Steam Next Fest October 2025 Sees Over 500 Demos Incorporate Generative AI, Sparking Industry Disappointment
The study serves as a critical reminder for both media organizations and the public to approach AI-generated information with caution and to prioritize verification through established, authoritative sources.
