After the errors committed by the generation of historical images by Google Gemini, the company's CEO, Sundar Pichai, defines the incident as "unacceptable".
"We have always strived to provide users with useful, accurate and unbiased information in our products.
That's why people trust us.
This must be our approach for all projects, including artificial intelligence ones", explained Pichai in an internal note, also taken up by the Wall Street Journal, in which he stated that the teams are working "24 hours a day" to solve the problem".
Gemini is a competitor to Microsoft Copilot, ChatGpt, and Dall-e, and other image-generative AI tools including Midjourney and Stable Diffusion.
The controversy over image generation emerged last week after many users discovered that Google's AI created historically inaccurate images that were biased towards people of white ethnicity.
"The system erred by being too inclusive," Prabhakar Raghavan, senior vice president of Google, admitted in recent days.
In the note, Pichai reported that he was aware that some of Gemini's responses "offended our users and showed bias" and added: "To be clear, this is completely unacceptable and we got it wrong."
Pichai said teams are "working around the clock to resolve these issues" and that the company has seen "substantial improvement," without going into further detail.
According to the CEO, the company will move forward with its AI roadmap based on announcements made in recent weeks.
Reproduction reserved © Copyright ANSA