Sergey Brin, co-founder of Google, has openly acknowledged the challenges faced by the company’s Gemini AI chatbot, specifically addressing issues surrounding its image generation capabilities. Brin admitted that Google had “messed up” with the rollout of this feature, leading to public criticism over historically inaccurate images generated by the chatbot.
Gemini, formerly known as Bard, had introduced text-to-image capabilities just last month. However, due to backlash over the accuracy of the images produced, Google decided to pause this feature temporarily.
Brin’s comments came during a rare public appearance at the AGI House in Hillsborough, California, where he engaged with developers and founders involved in testing Gemini. Responding to questions about the chatbot’s image generation problems, Brin attributed the issues to insufficient testing, expressing surprise at discussing the matter at the event.
Moreover, Brin highlighted that Google hasn’t fully understood why the Gemini chatbot tends to display a left-leaning bias in many cases. He clarified that such bias isn’t intentional, and the company has been actively working on improving accuracy.
Earlier, Google CEO Sundar Pichai had also addressed the problems with Gemini in a letter to the company’s employees, denouncing biased and offensive responses from the chatbot as completely unacceptable. Pichai emphasized the importance of rectifying these issues and ensuring unbiased and accurate interactions with users.
Purus ut praesent facilisi dictumst sollicitudin cubilia ridiculus.