Google announced on Thursday that it is temporarily halting its Gemini artificial intelligence image generation feature following reports of “inaccuracies” in historical depictions. Social media users raised concerns about the AI tool producing images of historical figures, such as the U.S. Founding Fathers, with inaccurate racial representations.
In a post on X on Wednesday, Google acknowledged that while the AI feature can “generate a wide range of people,” which is generally beneficial for users worldwide, it admitted that it was “missing the mark” in this instance. The tech giant assured that it is actively working to enhance the accuracy of such depictions.
An updated statement released on Thursday confirmed that Google will suspend the feature within Gemini responsible for generating images of people. The company intends to reintroduce the feature’s “improved” version soon.
The image generator tool debuted at the beginning of February under the Gemini platform, previously known as Bard. Its rollout faces challenges amid Google’s efforts to keep pace with Microsoft-backed OpenAI.
A CNBC reporter found that Gemini failed to produce any images when tested on Thursday morning. While Google grapples with issues surrounding Gemini’s image creation, OpenAI recently introduced Sora, its latest generative AI model capable of generating video content from text prompts.
Jack Krawczyk, a senior director of products for Gemini at Google, emphasized the company’s commitment to reflecting its “global user base” in image generation and seriously addressing concerns regarding representation and bias.
He noted that while the AI is adept at responding to open-ended prompts like images of people walking dogs, historical contexts require more nuanced handling, and adjustments will be made accordingly.
Google unveiled Gemini, its most advanced AI model, in late 2023. This month, it rebranded Bard, its ChatGPT competitor, as Gemini, and introduced a subscription service for a more potent version of the AI model.
