AI Assistants Misrepresent News Content in Nearly Half of Cases, Major Study Finds

A comprehensive international study has revealed that four leading AI chatbots misrepresent news content 45% of the time, raising serious concerns about their reliability as information sources. The unprecedented research, conducted by 22 public service media organizations including Deutsche Welle (DW), evaluated thousands of responses across multiple languages and regions.

Journalists from prominent broadcasters including the BBC and NPR assessed ChatGPT, Microsoft’s Copilot, Google’s Gemini, and Perplexity AI against criteria such as accuracy, sourcing, contextual information, and ability to distinguish between fact and opinion.

The findings paint a troubling picture of AI reliability. Almost half of all answers contained at least one significant issue, while 31% had serious sourcing problems and 20% included major factual errors. DW’s specific analysis found that 53% of AI responses to its questions contained significant issues, with 29% showing specific accuracy problems.

Among the factual errors discovered were AI systems naming Olaf Scholz as the current German Chancellor—despite Friedrich Merz having assumed the role a month earlier—and incorrectly identifying Jens Stoltenberg as NATO secretary general after Mark Rutte had already taken over the position.

“This research conclusively shows that these failings are not isolated incidents,” said Jean Philip De Tender, deputy director general of the European Broadcasting Union (EBU), which coordinated the study. “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.”

The findings are particularly concerning given the growing popularity of AI assistants as news sources. According to the Reuters Institute’s Digital News Report 2025, 7% of online news consumers already use AI chatbots to access news content, with that figure rising to 15% among users under 25.

The current study builds on previous research conducted by the BBC in February 2025, which found similar issues. In that earlier examination, more than half of AI-generated answers contained significant problems, and nearly one-fifth of responses citing BBC content introduced factual errors not present in the original material.

For the new study, media organizations from 18 countries across multiple language groups applied the same methodology to evaluate 3,000 AI responses. Journalists posed common news questions such as “What is the Ukraine minerals deal?” or “Can Trump run for a third term?” and then assessed responses based on their professional expertise without knowing which AI assistant had provided each answer.

While the results showed minor improvement compared to the BBC’s earlier findings, error rates remain troublingly high. Google’s Gemini performed worst among the four systems, with 72% of its responses containing significant sourcing issues.

“We’re excited about AI and how it can help us bring even more value to audiences,” noted Peter Archer, BBC program director of generative AI. “But people must be able to trust what they read, watch and see. Despite some improvements, it’s clear that there are still significant issues with these assistants.”

In response to these findings, the broadcasters are calling for government intervention. The EBU stated that its members are “pressing EU and national regulators to enforce existing laws on information integrity, digital services, and media pluralism.” They also emphasized the need for ongoing independent monitoring of AI systems, especially given how rapidly new models are being deployed.

Additionally, the EBU has joined with other international broadcasting groups to launch a campaign called “Facts In: Facts Out,” which demands that AI companies take greater responsibility for how their products handle news content.

“When these systems distort, misattribute or decontextualize trusted news, they undermine public trust,” the campaign organizers stated. “This campaign’s demand is simple: If facts go in, facts must come out. AI tools must not compromise the integrity of the news they use.”