Key study finding: AI systems like GPT-4 and Gemini Pro overestimated Western moral values while undervaluing non-Western perspectives in a large-scale comparison with human survey data. Why it ...