Google has announced that it will be allowing users to create images of people using its Gemini artificial intelligence tool after pulling the AI image-generation tool in February due to inaccuracies in historical pictures. Early access to Google’s new Imagen 3 generator will be available to Gemini Advanced, Business, and Enterprise users in English. The company has made technical improvements to the product, evaluated sets, and conducted red-teaming exercises to test for vulnerabilities.
The decision to pause the image generation feature earlier this year came after embarrassing examples circulated on social media, including racially diverse images when users asked for historical depictions. Imagen 3 will not support photorealistic identifiable individuals, depictions of minors, or excessively gory, violent, or sexual scenes.
Dave Citron, a senior director of product on Gemini, mentioned that although not every image generated by Gemini will be perfect, they will continue to listen to feedback from early users to improve the tool further. The gradual rollout of Imagen 3 will aim to bring it to more users and languages in the future. Overall, Google is working to enhance the AI tool and provide a better experience for its users.
Photo credit
www.nbcnews.com