Google has announced a temporary pause in Gemini’s ability to generate images of people using artificial intelligence. The company cited concerns over the portrayal of different ethnicities, genders, and historical figures, stating that a better version will be released soon.
In a post on X, formerly Twitter, Google acknowledged that the AI feature is capable of generating a diverse range of people, which is generally positive for users worldwide. However, the company admitted that the current software is missing the mark in some aspects and is actively working to enhance these depictions.
Gemini, previously known as Bard, introduced the image generator tool earlier this month. The decision to halt the production of images of people comes amidst criticism for creating inaccurate representations, such as images of historical figures like the US founding fathers depicted as American Indians.
Users accessing the feature are now informed that Gemini’s ability to generate images of people is being improved and will be reinstated soon. Google aims to notify users once the feature is updated.
The move by Google coincides with efforts to catch up with Microsoft-backed OpenAI, which recently unveiled Sora, a new generative AI model capable of producing videos from text prompts. The decision to pause Gemini’s image generation feature reflects an ongoing challenge in the development of accurate and unbiased AI models.
The University of Washington, Carnegie Mellon University, and Xi’an Jiaotong University conducted research in August highlighting potential political biases in AI models based on their development processes. Google’s past missteps, such as mislabeling photos in its app, underscore the importance of addressing diversity and accuracy in AI technologies.
As the tech giant works towards improving the portrayal of individuals in AI-generated images, the industry continues to grapple with ethical considerations and the need for inclusive representation in artificial intelligence.
