Google launches custom Gems and enhanced image generation with Imagen 3 in Gemini

Creating a custom Gem | Image credit — Google

High-quality image generation with Imagen 3

In addition to Gems, Google has also upgraded Gemini’s creative image generation capabilities with the introduction of Imagen 3. This new model is capable of generating high-quality images with just a few words and allows users to request images in different styles. Google emphasizes that Imagen 3 adheres to their product design principles and includes built-in safeguards. They have also implemented SynthID, their tool for watermarking AI-generated images.

A comparison of an Imagen 2 generated image (left) versus the same image generated on Imagen 3 (right) | Images credit — Google

One of the most notable advancements with Imagen 3 is the ability to generate images of people. However, Google has taken a cautious approach with this feature, initially offering an early access version for Gemini Advanced, Business, and Enterprise users in English. They have made technical improvements, refined evaluation sets, and established clear product principles to ensure a responsible and safe user experience. While they acknowledge that not every image generated will be perfect, they are committed to continuous improvement based on user feedback.

Sample Imagen 3 images and prompts | Images credit — Google

These updates highlight Google’s ongoing efforts to enhance Gemini and provide users with a more personalized and powerful AI experience. The introduction of Gems and Imagen 3 demonstrates Google’s commitment to innovation and its dedication to pushing the boundaries of what AI can achieve. Google encourages users to try Gemini Advanced or sign up for Gemini for Workspace to experience these new features firsthand.