Google is pausing an AI tool that creates images of people following inaccuracies in some historical depictions the model generated. (Gonzalo Fuentes/Reuters)/ CBC News
Google has decided to temporarily halt the use of an AI tool designed to generate images of people due to inaccuracies found in some historical depictions created by the model. This move marks the latest setback in the efforts of the Alphabet-owned company to keep pace with competitors like OpenAI and Microsoft in the field of artificial intelligence.
Earlier this month, Google introduced image generation capabilities through its Gemini AI models. However, concerns arose over the past few days as users on social media pointed out instances where the model produced historical images that were inaccurate. Examples included depictions of U.S. Founding Fathers as women or Nazi soldiers portrayed as individuals of color.
In response to these concerns, Google issued a statement on Wednesday acknowledging the inaccuracies in some historical image depictions generated by Gemini.
Google has been striving to develop AI software that can rival the capabilities of OpenAI's ChatGPT, especially since the launch of ChatGPT in November 2022. However, Google has faced challenges along the way. Last year, the company's generative AI chatbot, Bard, drew criticism for sharing inaccurate information about pictures of a planet beyond Earth's solar system in a promotional video. This incident led to a decline in Google's shares by as much as nine percent.
Recently, Bard was rebranded as Gemini, and Google introduced paid subscription plans offering users access to a more advanced AI model. Jack Krawczyk, senior director of product for Gemini at Google, stated on Wednesday that historical contexts present more complexity, and Google aims to fine-tune its AI models to better accommodate such nuances.
