Chinese researchers found that ChatGPT is experiencing a rightward shift in political values, despite still being classified as 'libertarian left'. The study involved testing various ChatGPT models with the Political Compass Test and revealed a significant rightward tilt in responses over time. This shift may be influenced by changes in training datasets, user interactions, or updates to the chatbot. The researchers emphasize the need for continuous scrutiny of AI models to prevent the spread of skewed information and societal polarization. Teljes cikk (Euronews.com)