Artificial Intelligence Dramatically Reshapes News Media: BBC and EBU Study Reveals Concerning Trends
According to a comprehensive international study conducted by the European Broadcasting Union (EBU) based on extensive BBC data, alarming signals emerge about the impact of artificial intelligence (AI) on the objectivity and reliability of modern news.
During the News Assembly in Naples, over 3,000 responses from leading AI assistants such as ChatGPT, Copilot, Gemini, and Perplexity were analyzed, revealing that in 45% of cases, these systems distort or misrepresent news content.
This poses a significant threat to public trust in media.
The study identified serious issues with sources—31% of responses lacked proper sourcing or contained incorrect references, while 20% included factual errors or outdated information.
The platform Gemini showed the worst results, with over 76% of responses being inaccurate—more than twice as bad as other assistants.
Jean-Philippe Tenders, EBU Media Director and Deputy General Director, emphasized that such systemic problems threaten the credibility of media institutions and raise concerns about information security globally.
It is estimated that over 7% of online news consumers already receive information via AI assistants, with the figure rising to 15% among those under 25.
BBC’s Generative AI program highlights ongoing challenges related to quality and ethical use of automated systems.
In response, a “News Integrity Toolkit for AI Assistants” was developed to help developers improve answer accuracy and enhance media literacy among users.
Additionally, EBU calls for stricter regulation of digital information and media pluralism in Europe and beyond.
A Graphite study notes that more than half of all written online content is now generated by AI, a figure that surged after the November 2022 launch of ChatGPT, even though search engines like Google remain effective at detecting AI-generated content—only 14% of search results are AI-produced.