A recent study by the BBC found that four major generative AI chatbots, including OpenAI’s ChatGPT and Microsoft’s Copilot, provided “significant inaccuracies” and distortions when summarizing news. The organization tested the services by feeding them content from its website and then asking them questions about the news.
The results showed that 51% of all AI-generated answers had flaws, with 19% introducing errors including incorrect statements of facts, numbers, and dates. Examples of mistakes included Gemini incorrectly stating that the UK’s National Health Service did not recommend vaping as a quit smoking aid, while ChatGPT and Copilot wrongly stated that politicians Rishi Sunak and Nicola Sturgeon were still in office.
BBC News and Current Affairs CEO Deborah Turness called on tech giants to improve the quality of AI-powered news tools, warning that distorted content can cause significant real-world harm. The study’s findings come after Apple recently suspended its AI-powered news alerts following reports of inaccurate content.
Turness praised Apple’s decision, stating that “the companies developing Gen AI tools are playing with fire.” She emphasized the importance of ensuring accurate and reliable information, particularly in a world where people often seek clarity in chaotic times.
Source: https://deadline.com/2025/02/ai-chatbots-inaccurate-news-summaries-bbc-tech-1236285723