Google CEO Sundar Pichai Admits Lacks And Bias In Gemini

JAKARTA - Google CEO, Sundar Pichai, told employees in a note on Tuesday 27 February that the company was working to repair its AI Gemini tool. He called some text and image responses generated by the model "bias" and "absolutely unacceptable".

The company last week stopped using its tools to image people after the inaccuracies in some of the historical descriptions generated.

Pichai told employees that some of the responses to the tool had offended its users and showed bias.

"Our team has been working day and night to solve this problem. We have seen a significant increase in various prompts... And we will review what happened and make sure we fix it widely," he said.

Google is now planning to re-launch Gemini AI in the coming weeks. The Semafor news site first reported the news, which was later confirmed by a Google spokesperson.

Since the launch of ChatGPT by Microsoft-backed OpenAI in November 2022, Alphabet's Google has been racing to create competitor AI software.

The company released the general AI chatbot Bard a year ago. Earlier this month Google changed its name to Gemini and launched a paid subscription plan, which users can choose for better reasoning capabilities than the AI model.