"We Definitely Messed Up": Google Co-founder Admits Errors In Gemini AI

Google co-founder Sergey Brin recently acknowledged that the tech giant's AI model, Gemini, is a "work in progress" and openly admitted to errors in the image generation aspect of Gemini.

In a video, recorded at San Francisco's AGI House, he can be heard saying, "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.”

Sergey Brin at the AGI House.

Must watch pic.twitter.com/IofAUYdGvA

Gemini came under fire recently for generating historically inaccurate images, including racially diverse representations of Nazis. 

According to Mr Brin, 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, as reported by some users, reported Fortune. 

Mr Brin added that he “kind of came out of retirement just because the trajectory of AI is so exciting.” He has also acknowledged that the technology is still a work in progress, the report added.

He also compared Gemini's errors to potential issues in other large language models. 

“If you deeply test any text model out there, whether it's ours, ChatGPT, Grok, what have you, it'll say some pretty weird things that are out there that you know definitely feel far left,”  Mr Brin was quoted as saying. 

He mentioned the company is uncertain why Gemini tends to "lean left in many cases" but assured it is not intentional, reported CNBC. 

“We haven't fully understood why it leans left in many cases but if you try it starting over this last week, it should be at least 80% better, of the test cases that we've covered,” he was quoted as saying.  

However, despite the setbacks, Mr Brin remains optimistic about the future of AI, expressing excitement and involvement in writing code, the Fortune report added.