- Google announced on Thursday that it is temporarily halting the use of its Gemini artificial intelligence image generation feature due to identified “inaccuracies” in historical images.
- Complaints from users on social media highlighted concerns about the AI tool depicting historical figures, such as the U.S. Founding Fathers, as individuals of color, which was deemed as inaccurate.
- An updated statement from Google on Thursday mentioned the suspension of Gemini’s image generation feature, with plans to introduce an “enhanced” version in the near future.
Google revealed on Thursday its decision to pause the implementation of the Gemini artificial intelligence image generation tool following reports of “inaccuracies” in historical image outputs.
There have been reports on social media platforms citing dissatisfaction with the AI tool’s rendering of historical figures, including iconic personalities like the U.S. Founding Fathers, as individuals of different racial backgrounds, which was perceived as historically inaccurate.
In a recent post, Google acknowledged the broad spectrum of images that the AI feature is capable of generating, considering it a positive aspect widely utilized by users globally. However, it recognized the shortcomings in this particular aspect, emphasizing its commitment to rectifying these errors promptly.
Furthermore, Google released an updated statement on Thursday confirming the temporary suspension of Gemini’s image generation feature for depictions of individuals, promising an improved version to be reintroduced soon.
The image generation tool was initially launched in early February under the Gemini project, previously known as Bard. This development comes amidst challenges for Google as it strives to keep pace with the Microsoft-supported OpenAI.
During a test conducted by a CNBC reporter on Thursday morning, Gemini failed to produce any images at all, indicating technical difficulties.
While Google grapples with issues related to Gemini’s image creation functionality, OpenAI recently unveiled Sora, its latest generative AI model capable of generating videos based on text prompts from users.
Jack Krawczyk, Google’s senior director of product for Gemini, emphasized the company’s dedication to reflecting the diversity of its global user base through its image generation capabilities, underscoring its commitment to addressing representation and bias concerns seriously.
Krawczyk stated in a recent communication that while the AI model excels in responding to open-ended prompts like “images of a person walking a dog,” it recognizes the need for further refinement to accurately capture the complexities of historical contexts.
Google introduced Gemini, its cutting-edge AI model, towards the end of 2023. Additionally, this month witnessed the rebranding of Bard, Google’s ChatGPT competitor, as Gemini, alongside the launch of a subscription service offering access to an enhanced version of the AI model.
