Google CEO Sundar Pichai said recent “problematic” text and image responses from the company’s Gemini artificial intelligence (AI) chatbot were “completely unacceptable.”
“I know that some of its responses have offended our users and shown bias – to be clear, that’s completely unacceptable and we got it wrong,” Pichai said in an email to employees Tuesday, first reported by Semafor.
The company has been “working around the clock to address these issues” and is “already seeing substantial improvement,” he added.
“No AI is perfect, especially at this emerging stage of the industry’s development, but we know the bar is high for us and we will keep at it for however long it takes,” Pichai continued. “And we’ll review what happened and make sure we fix it at scale.”
Google paused Gemini’s ability to generate images last week after the chatbot produced “inaccuracies in some historical image generation depictions,” including racially diverse Nazi soldiers and U.S. Founding Fathers.
Prabhakar Raghavan, a senior vice president at Google, explained in a blog post Friday that the company’s efforts to ensure that Gemini showed a range of people “failed to account for cases that should clearly not show a range.”
The AI model also became “way more cautious” than intended, refusing to answer certain prompts entirely, Raghavan noted.
“These two things led the model to overcompensate in some cases, and be over-conservative in others, leading to images that were embarrassing and wrong,” he said.
“It’s clear that this feature missed the mark,” Raghavan added in the blog post.
Pichai said in Tuesday’s note that Google will be “driving a clear set of actions, including structural changes, updated product guidelines, improved launch processes, robust evals and red-teaming, and technical recommendations” in response to the missteps with Gemini.