Research led by the BBC has revealed that prominent AI tools, including ChatGPT, Copilot, Gemini, and Perplexity, deliver answers about news and current events that often contain factual errors and misrepresentations. Over half of the AI-generated responses in the study had significant issues, including outdated claims about political figures and inaccurate interpretations of health advice. The analysis involved 100 questions based on BBC articles, with many answers judged poorly by BBC journalists, highlighting the urgent need for improved accuracy in AI communications.
Leading artificial intelligence assistants often produce significant distortions, inaccuracies, and misleading content, particularly in responses related to news and current affairs.
Collection
[
|
...
]