Summary:
- Researchers at the University of Hong Kong (HKU) evaluated several Chinese-developed AI language models and found that they struggle with "hallucinations" - generating text that appears plausible but is factually incorrect.
- The study found that these AI models often produce responses that are coherent but not grounded in reality, which could be problematic for real-world applications.
- The researchers suggest that more work is needed to improve the reliability and trustworthiness of Chinese AI language models, especially as they become more widely used in various industries.