ChatGPT.Gov: AI for US Government Agencies

Описание: We dive deep into a Euronews article that reports on a study that examined ChatGPT's political leanings over time. Researchers from Peking University used the Political Compass Test to analyse ChatGPT models. The study found a "rightward tilt" in how the AI answered questions, even though it still maintained "libertarian left" values. Researchers suggested the shift could be due to changes in training data, user interactions, or updates to the chatbot itself. The article stresses the need for constant monitoring and transparency in AI models to ensure they are fair and balanced, preventing the spread of skewed information.