Google will soon restore access to the image generator in its AI chatbot Gemini. The feature was removed in February due to serious historical errors in depicting people, particularly related to race and gender. For instance, the racial diversity shown in images of a “Roman legion” was a clear anachronism.
Improvements and Limited Access
Dave Citron, senior director of products for Gemini, announced on the official Google blog that early access to Google’s new image generator, Imagen 3, will be available to paid Gemini users on Advanced, Business, and Enterprise plans in the coming days. Initially, the feature will only support queries in English.
“We made technical fixes to the product, as well as worked on more advanced rating algorithms and protection against red-teaming attacks,” Citron wrote. The company decided to revise the Gemini image generator less than a day after receiving the first complaints about historical inaccuracies.
According to Citron, the new Imagen 3 generator “does not support the creation of photorealistic identifiable faces, images of minors, or excessively bloody, violent, or sexual scenes.” He acknowledged that not every image created by Gemini will be perfect, but assured users that Google will continue to listen to feedback and improve the product.
Future Plans and Limitations
Citron promised that more people will be able to use the image generator in the future, and the feature itself will receive support for additional languages. We’ll keep you updated on these developments as they occur.
NIX Solutions notes that while Google is working to address the issues that led to the feature’s removal, challenges remain in creating historically accurate and diverse representations. The company’s approach of limiting certain types of image generation, such as photorealistic faces and sensitive content, demonstrates their cautious stance in reintroducing this technology.
As AI-generated imagery becomes more prevalent, it’s crucial for users to remain aware of potential inaccuracies and biases that may still exist in these systems. Google’s ongoing efforts to improve Gemini’s image generation capabilities reflect the broader challenges faced by AI companies in balancing creativity, accuracy, and ethical considerations.