Google co-founder Sergey Brin has admitted to errors in the image generation aspect of the company’s AI model Gemini. He has said that Gemini is a “work in progress”.
“We definitely messed up on the image generation. I think it was mostly due to just not thorough testing. It definitely, for good reasons, upset a lot of people,” he said in a video.
Gemini was in the midst of a controversy recently for generating historically inaccurate images, including racially diverse representations of Nazis.
Media reports quote Brin as saying that the algorithm’s unintentional bias resulted in inaccurate, non-white images for prompts related to figures such as Adolf Hitler, the Pope, and medieval Viking warriors.
Brin, who said he is excited by the “trajectory of AI” compared Gemini’s errors to potential issues in other large language models. If any text model is tested deeply some “pretty weird things” could come up, he added.
He acknowledged that the company had not understood why Gemini tends to “lean left in many cases” but said this was not intentional. Despite the setbacks, he said he is optimistic about the future of AI.
Also Read: India Powers Up: Global Semiconductor Ambitions in 5 Years
