Google CEO Sundar Pichai admits shortcomings in Gemini, the AI feature meant for generating images from user prompts.
In a recent memo, Pichai labeled Gemini’s errors as ‘problematic’ and ‘unacceptable’, citing offense and bias among users.
Users highlighted historical inaccuracies in Gemini-generated images, leading to viral attention on the issues.
As a result, Google has temporarily taken the feature offline for further testing and plans to re-launch it in the coming weeks.
Pichai stressed meeting high standards in AI development, acknowledging Google’s mistake given the early stage of AI development.