In a bid to enhance diversity and inclusion, Google's Gemini AI chatbot faced criticism for inaccuracies in image generation. The feature launched three weeks ago, prompted users to question its depiction of historical figures, accusing it of oversaturating diversity.
Acknowledging the concerns, Google Senior Vice President Prabhakar Raghavan issued a statement, stating, "Gemini's image generation feature missed the mark, and we're grateful for users' feedback." The company temporarily halted the feature to address inaccuracies and offensive depictions.
Raghavan emphasized the intention behind Gemini's image generation, which aimed to counter issues seen in other products, such as the inappropriate use of AI for violent or explicit content. Google incorporated diversity, equity, and inclusion standards but admitted to overshooting the mark.
"We've acknowledged the mistake and temporarily paused image generation of people in Gemini while we work on an improved version," Raghavan added.
Gemini's image generation feature utilized the Imagen 2 AI model. While seeking to represent diverse scenarios accurately, the AI reportedly struggled with specific prompts, depicting inaccuracies in responses.
Google assured users of extensive work, including rigorous testing, before reintroducing the improved image generation feature. The temporary pause serves as a commitment to rectifying issues and delivering a more refined and accurate Gemini AI experience.
0 Comments