The news: The BBC found that over half of news summaries generated by major large language models (LLMs) contained significant problems, including misquotes, editorial bias, and outdated information.
It’s the latest setback for Big Tech platforms and solutions leaning on AI features to push user adoption.
Can’t trust AI news summaries: The study examined OpenAI’s GPT-4o, Microsoft Copilot Pro, Google Gemini Standard, and Perplexity. Here are some findings.