Gemini AI Controversy: Google Explains Why Image Generation Tool 'Missed The Mark'; All You Need To Know

Google infused image generation capabilities in Gemini AI with Bard chatbot's rename a few weeks ago. This included the option to generate images of people. However, as the issues of "inaccurate historical" images came to light, Google has now admitted that the "feature missed the mark" and some output generated from the model was "inaccurate or even offensive". While acknowledging the error, the company has now temporarily paused the photo generation of people in Gemini for further improvements.

Sharing the update in a Google blog, Prabhakar Raghavan, Senior Vice President, said that the model is powered by Imagen 2 and was tuned to serve a range of choices to users who prompted Gemini for image generation. However, Raghavan added that the tuning "to ensure that Gemini showed a range of people failed to account for cases that should clearly not show a range."

Also Read: 'Direct Violations': IT Minister Reacts After X User Flags Google Gemini's Response On PM Modi

Moreover, Gemini took extra caution when answering queries and refused to answer some prompts - "wrongly interpreting some very anodyne prompts as sensitive". The two cases led to "overcompensate in some cases [output of racially diverse Nazis], and be over-conservative in others, leading to images that were embarrassing and wrong."

Google has promised significant improvements and extensive testing ahead of making Gemini's image generation available again. (Image:Unsplash)

While the image generation of individuals is now restricted, Google has promised significant improvements and extensive testing ahead of making the feature available again. However, Google still does not promise reliable results for the image and text generation capabilities. "It will make mistakes", as hallucinations are a challenge to large language models (LLMs).

Also Read: Google Halts Gemini AI's Feature To Create AI Images Of Individuals; Here's Why

"I can't promise that Gemini won't occasionally generate embarrassing, inaccurate or offensive results — but I can promise that we will continue to take action whenever we identify an issue," Raghavan added. Google also emphasised the double-check feature, which aims to give factual information as a response to a prompt in Gemini. Meanwhile, this update also comes hours after IT Minister Rajeev Chandrasekhar flagged Gemini's response as"direct violation" of several Indian laws. You can read more on this here.