Tech Desk 
Oct 23, 2025
An AI (Artificial Intelligence) sign is seen at the World Artificial Intelligence Conference (WAIC) in Shanghai, China July 6, 2023. — Reuters
While the impression that AI is taking over the roles of news writers and editors has been widely evident, a recent study has revealed that nearly half of the responses generated by leading AI assistants misrepresent news content.
The research, conducted by the European Broadcasting Union (EBU) and the BBC, analysed 3,000 responses from AI systems, including ChatGPT, Copilot, Gemini, and Perplexity, across 14 languages, with the focus on accuracy, sourcing, and the ability to distinguish between facts and opinions.
The results of the study on AI for news content were very disconcerting. It showed that 45% of AI responses contained significant issues, while 81% had some form of problem.
The most noteworthy finding was that one-third of the responses had serious sourcing errors, and 72% of Gemini’s outputs showed significant issues compared to under 25% for others. Besides, 20% of the responses contained factual errors, including outdated information and misstatements about current events.
The EBU warned that as AI assistants increasingly replace traditional search engines for news, public trust could be at risk. EBU Media Director Jean Philip De Tender highlighted the potential consequences, stating, “When people don’t know what to trust, they end up trusting nothing at all, which can deter democratic participation.”
It was also observed that as of 2025, 7% of online news consumers and 15% of users under 25 rely on AI for news.
The report urges holding AI developers accountable and improving how their platforms handle news-related queries to ensure accuracy and reliability.
