Google has announced a temporary halt to its Gemini AI chatbot’s image generation feature after facing criticism for inaccuracies and racial biases in its historical depictions, Associated Press reports.
Users shared screenshots of racially diverse characters in predominantly white scenes, raising concerns about over-correction for racial bias. Previous studies have shown AI image generators can perpetuate racial and gender stereotypes.
Google acknowledged the inaccuracies and pledged to improve the depictions. Some researchers are conflicted about the response, emphasizing the complexity of addressing representational harm in AI systems.
