According to a BBC investigation, AI chatbots from major tech firms are struggling with accuracy when summarizing news. The study found that OpenAI’s ChatGPT, Microsoft’s Copilot, Google’s Gemini, and Perplexity AI frequently misrepresented facts, introduced errors, and failed to provide proper context in their summaries.

The BBC tested these AI systems by providing them with news articles and analyzing their responses. The results were concerning: 51% of reactions contained significant issues, while nearly 20% of answers citing BBC content included factual errors such as incorrect dates, numbers, and misattributed quotes. Some examples included Gemini falsely claiming the NHS does not recommend vaping as a smoking cessation aid, ChatGPT and Copilot failing to recognize leadership changes in the UK, and Perplexity misrepresenting BBC reports on Middle East conflicts.

BBC News and Current Affairs CEO Deborah Turness criticized these inaccuracies, warning that AI-generated misinformation could have real-world consequences. She urged AI developers to “pull back” on news summarization features, echoing concerns that led Apple to suspend its AI-generated news alerts after similar complaints.

The report highlighted broader issues with AI-generated content, including the chatbots’ inability to distinguish between fact and opinion, a tendency to editorialize, and a lack of essential context. Microsoft’s Copilot and Google’s Gemini were found to have the most serious accuracy problems, while ChatGPT and Perplexity performed slightly better but still exhibited notable flaws.

The BBC, which generally blocks AI bots from scraping its content, made an exception for this study to assess how these models process news. Pete Archer, the BBC’s Programme Director for Generative AI, stressed that publishers should have control over their content and that AI companies must be transparent about generating news summaries.

This latest controversy adds to growing concerns over the role of AI in information dissemination, raising urgent questions about accountability, misinformation, and the ethical responsibilities of tech companies deploying AI-powered tools.