Google to Bring Back Gemini AI Image Tool Soon

Google's Gemini might make a comeback after the temporary discontinuation of its AI image-generation feature.

Google DeepMind CEO Demis Hassabis said that they are working on the shortcomings of Gemini's photo feature for better accuracy and will be relaunched within a few weeks. He also said that the AI tool was not working the way it was intended.

 Google Gemini AI was temporarily discontinued for depicting inaccurate historical images as in specific white figures like the US Founding Fathers, and Nazi-era German soldiers as people of color. Google's Gemini faced a major setback for its inaccurate images creating racial bias.

Previously, Google's Vice President, Prabhakar Raghavan, acknowledged that Gemini's image generation feature "missed the mark" and generated inaccurate images that led to racial bias. He also apologized for the shortcomings of Gemini.

He also said that a few issues occurred in the Gemini image generation feature. This involves first the tuning process failed to account for a range of diversity in images that seemed inappropriate. Later, over time the model became excessively cautious and refused to respond to certain prompts incorrectly by deeming innocuous queries as sensitive.

Gemini AI was officially released on December 6 by Google. As it integrates natural language processing and image recognition, Gemini helps in image captioning and complex visual parsing. It is not the first time that AI has stumbled into questions over racial bias and diversity. Even earlier about a decade ago Google had to apologize when the photo app conveyed a black couple as gorillas.