Google will temporarily halt Gemini AI model’s creation of images of people due to inaccuracies
Google has decided to temporarily halt its AI tool that generates images of people due to inaccuracies in historical depictions produced by the model. This setback is the most recent in the company’s attempts to compete with rivals OpenAI and Microsoft.
Google started generating images through its Gemini AI models earlier this month, but in recent days some social media users had reported that the model returns historical images that are sometimes inaccurate.
“We are aware that Gemini provides inaccuracies in some historical image generation descriptions,” Google said Wednesday.
Since the launch of OpenAI’s ChatGPT in November 2022, Google has been racing to produce AI software that rivals the Microsoft-backed company.
When Google launched Bard, a generative AI chatbot, a year ago, the company had misrepresented images of a planet outside Earth’s solar system in a promotional video, causing shares to slide as much as 9%.
Bard was named Gemini earlier this month, and Google introduced paid subscription plans that users could choose to access the AI model’s higher reasoning capabilities.
“Historical contexts are more nuanced, and we’re still tweaking them to reflect that,” said Jack Krawczyk, Google’s Gemini product director, on Wednesday.
