Major AI chatbots like ChatGPT, Google's Gemini, Perplexity, and Copilot misrepresent news content in 45% of their responses, new study revealed. AI assistants were assessed in 14 languages for accuracy, sourcing and ability to distinguish opinions from facts. Around 72% of responses by Gemini had sourcing issues, while accuracy issues were found in 20% of responses from all AI bots.