Google apologizes for AI’s error on Founding Fathers, halts feature

According to Global Times citing the report of Associated Press on the 24th, Prabhakar Raghavan, the Senior Vice President of Google, apologized on the 23rd for the issues with the AI model Gemini, developed by the company, in generating person images. Combining reports from Associated Press and other media outlets, Google’s AI model Gemini is designed to generate images based on text descriptions. However, some users recently noticed that most generated person images are of people of color rather than white individuals. Some users believe that the model has a “anti-white” bias in the image generation process. In response, Google stated that they are taking measures to temporarily halt the function of generating person images by Gemini. Raghavan expressed in a post on the 23rd, “It is obvious that this function did not deliver the expected results.” He also mentioned, “Some of the generated images are inaccurate or even unpleasant. We appreciate the feedback from users and apologize for the function not working as intended.” (Source: Dongfang, Video by Jiang Meiqing)

Leave a Reply

Your email address will not be published. Required fields are marked *

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.