AI Chatbots Misrepresent News Nearly 50% of the Time, Study Finds

A comprehensive study conducted by a global alliance of public broadcasters reveals that AI chatbots misrepresent news content nearly 50% of the time. This assessment involved 22 public media organizations across 18 countries, including Canada’s CBC/Radio-Canada. The research focused on how AI chatbots respond to inquiries about news and current affairs, highlighting significant accuracy concerns.

The study evaluated over 3,000 responses generated by four prominent AI chatbots: OpenAI’s ChatGPT, Microsoft’s Copilot, Google’s Gemini, and Perplexity. Journalists from various public media outlets scrutinized these answers and identified major issues in 45% of cases.

Findings on Accuracy and Sourcing

Among the critical findings, the study pointed out that sourcing problems afflicted 31% of the responses. In addition, 20% of the answers contained significant accuracy issues. These discrepancies raise concerns about the reliability of AI chatbots as sources of information, particularly in an era where they are increasingly replacing traditional search engines.

The European Broadcasting Union stated in a press release that the study underscores the challenges posed by AI technology in the news landscape. They noted that these chatbots “routinely misrepresent news content,” which could have implications for public understanding and trust in media.

The report, published on October 22, 2025, by The Canadian Press, emphasizes the need for caution when relying on AI-generated information. Anja Karadeglija, a journalist with The Canadian Press, highlighted the importance of verifying facts in an age where misinformation can spread rapidly through digital platforms.

As AI technology continues to evolve, the findings of this study serve as a timely reminder of the potential pitfalls associated with automated information dissemination. Public broadcasters and consumers alike are urged to remain vigilant and critically assess the reliability of AI-generated news content.

In conclusion, the study reflects a growing concern about the impact of AI on journalism and the dissemination of accurate information. With nearly half of the responses examined containing significant errors, the role of AI chatbots in informing the public remains a topic of considerable debate and scrutiny.