Google CEO Sundar Pichai isn’t happy with what’s happening with the company’s Gemini AI chatbot, and in an internal memo to employees on Tuesday he said the company’s image generator Viral and historically inaccurate errors were “unacceptable”.
“We want to address recent issues with problematic text and image responses in the Gemini app (formerly Bard),” Pichai said in the note. semaphores and later confirmed The Verge. “We know that some of those responses were offensive and showed bias. Let’s be clear: It’s completely unacceptable and we made the wrong decision.”
Pichai explained that Google’s team has been “working around the clock” to address the Gemini issue and has “seen significant improvements across a wide range of prompts.” Gemini’s image generator caused an uproar online last week after it was created by an AI chatbot. Historically inaccurate images of black Vikings and refused to produce images of white people.
Google immediately shut down Gemini’s portrait creation feature last Thursday in the wake of the scandal, and Google Vice President Prabhakar Raghavan apologized for the incident on Friday.
“This was not what we intended. We did not want to deny Gemini the ability to create images of certain groups, and we did not want to reject historical or other images that were inaccurate. Nor did we want to create one,” Raghavan said. blog post.
The company aims to reboot Gemini’s image generator over the next few weeks.
While Raghavan’s statement was aimed at the public, Pichai’s memo was aimed at employees, some of whom may have been surprised by Gemini’s announcement. Some people are Claims to work in Google’s AR & VR department Women in San Francisco say they have “never been more embarrassed to work for a company” after Gemini’s historically inaccurate image of a medieval English king went viral. Gizmodo has not been able to independently confirm whether this person works at Google.
In his memo, Pichai said the company would make structural changes to avoid a repeat of the situation with Gemini. He also stressed that it is important for Google to provide users with “useful, accurate and unbiased” information about all of its products, including emerging AI products.
“No AI is perfect, especially in this new phase of industry development, but we know the hurdles are high for us and we will keep trying no matter how long it takes. And what happened? We will look at it and make sure we fix it on a large scale,” Pichai said.