Google’s Gemini AI, touted as the most powerful chatbot available to the public, faced criticism for generating historically inaccurate images, including Black Nazi soldiers and Native American U.S. senators. The company announced a pause on Gemini’s image generation capabilities to address these inaccuracies. Former Google search engineer Darghya Das highlighted the bot’s failure to acknowledge white people, while right-wing accounts criticized the diverse images generated by Gemini. The issue stemmed from a training and labeling problem, where the AI overcorrected to include diversity in unrelated prompts, leading to racially diverse images of historical figures. Users also found instances where Gemini generated images of Black people but not white people, sparking accusations of racism against white individuals.
Despite being praised as an impressive chatbot, Gemini’s flaws underscore the challenges of bias in generative AI. The episode highlights the complexities of navigating cultural sensitivities and historical accuracy in AI technology. The company’s decision to pause image generation and work on improvements reflects the ongoing struggle to address bias and ensure accurate and culturally sensitive outputs. The incident serves as a cautionary tale of the unintended consequences of AI technology and the importance of addressing bias in machine learning algorithms.
Source link
Source link: https://ca.news.yahoo.com/google-pulls-geminis-ai-image-171024089.html