Google has unveiled two significant upgrades to its Gemini apps, expanding their capabilities with advanced artificial intelligence (AI) tools. The announcement, made on Wednesday, introduces the integration of “Gems,” a new AI agent, and the Imagen 3 image generation model. These enhancements were first previewed at the Google I/O earlier this year.
Gems: Advanced AI Chatbot Functionality
Gems, the new feature for Gemini, represents a notable advancement in AI chatbot technology. This tool allows users to create specialized versions of the chatbot tailored to specific topics or tasks. These miniature chatbots, with a limited dataset, are designed to deliver more precise and relevant responses.
Google elaborated on the benefits of Gems, stating, “With Gems, you can create a team of experts to help you think through a challenging project, brainstorm ideas for an upcoming event, or write the perfect caption for a social media post.” Gems can be customized with specific instructions to enhance their effectiveness. Users will have access to a range of pre-made Gems, including roles such as Learning Coach, Brainstormer, Career Guide, Writing Editor, and Coding Partner. This feature will be available to paid users, including those subscribed to Gemini Advanced, Business, and Enterprise tiers.
Imagen 3: Cutting-Edge Image Generation
The Imagen 3 model, Google’s latest image generation AI tool, is set to be integrated into Gemini apps, offering users enhanced capabilities for creating images. Imagen 3 can generate visuals in various styles, such as Nikon DSLR, GoPro, wide-angle lens, and more. The tool is designed to produce a wide range of outputs, including photorealistic landscapes, textured oil paintings, and whimsical claymation scenes.
A notable upgrade with Imagen 3 is its ability to generate images of people. This feature had been previously removed due to concerns about generating biased or harmful content. To address these issues, Google has implemented safeguards to reduce the risk of deepfakes and has introduced SynthID for watermarking AI-generated images.
Access and Limitations
The Imagen 3 capabilities will be rolled out to all Gemini users, including those on the free tier, though they might encounter certain limitations compared to paid users. The company hinted at the possibility of inline editing of generated images using text prompts, though specific details on this functionality were not provided. Importantly, Google has stated that Imagen 3 will not support the generation of photorealistic, identifiable individuals, depictions of minors, or excessively gory, violent, or sexual content.
These updates to Gemini apps are expected to enhance user experience by providing advanced tools for both communication and creative tasks.
