This deep dive discusses a new study revealing widespread errors in how popular AI assistants misrepresent news content, raising concerns about public confidence in information. The joint research, conducted by the European Broadcasting Union and the BBC, analyzed 3,000 responses from leading AI models like ChatGPT and Gemini, finding that nearly half of all responses to news queries contained significant issues. Specifically, the research found prevalent problems with sourcing errors, such as missing or incorrect attribution, and factual inaccuracies, including outdated or false information. The text notes that these systemic issues, exemplified by instances where AI misidentified political leaders or reported inaccurate legal details, could seriously undermine public trust and democratic participation. Consequently, media organizations are calling for AI companies to improve accuracy and accountability regarding how their products handle and redistribute news. You can read the source article here