Updated February 28th, 2024 at 16:57 IST

Google fixing issues in Gemini AI after CEO flags some responses ‘unacceptable’

Last week, Google paused the use of the Gemini AI tool following inaccuracies in historical depictions it generated.

Reported by: Business Desk
Gemini AI | Image:Google Gemini
Advertisement

Google fixing Gemini AI: Google CEO Sundar Pichai informed employees in a note on Tuesday that the company is actively addressing issues with its Gemini AI tool, acknowledging some of the generated text and image responses as "biased" and "completely unacceptable".

Last week, Google paused the use of the Gemini AI tool following inaccuracies in historical depictions it generated. Pichai expressed concern that some of the tool's responses had offended users and exhibited bias.

Advertisement

The Gemini AI-generated images included racially diverse Nazi-era German soldiers, non-white US Founding Fathers, and inaccuracies in depicting Google's own co-founders. While the company has apologised for missing the mark and is working to re-enable image generation, Pichai's memo is the first public acknowledgement of the controversy.

"Our teams have been working diligently to rectify these issues. We're already observing significant improvements across various prompts... We will thoroughly review the situation and ensure comprehensive fixes are implemented," Pichai stated. Google intends to reintroduce the Gemini AI tool in the coming weeks. The news was initially reported by the Semafor news website and later confirmed by a Google spokesperson.

Advertisement

Since the debut of Microsoft-backed OpenAI's ChatGPT in November 2022, Google, under Alphabet Inc., has been striving to develop a competing AI software. A year ago, it launched the generative AI chatbot Bard, which was recently rebranded as Gemini and accompanied by paid subscription plans offering better reasoning capabilities from the AI model.

(With Reuters inputs)

Advertisement

Published February 28th, 2024 at 15:09 IST