"Google's Gemini AI: Addressing Glitches and Diversity Controversies"

Google's Gemini AI tool generated "embarrassing and wrong" images due to tuning issues, including racially diverse Nazis and US Founding Fathers. Google's senior vice president, Prabhakar Raghavan, explained that the model overcompensated and became over-conservative, leading to inaccurate and cautious responses. Google has stopped allowing users to create images of people with Gemini AI and plans to improve its image-generation abilities before reenabling it. Raghavan acknowledged the challenge of "hallucinations" with large language models and emphasized ongoing efforts to improve accuracy.