Google announces Gems, Imagen 3, and an early preview for improved people-generation capabilities

Google recently shared that it would start rolling out two features previewed at Google I/O: Gems, the personalized AI-powered assistants, will be available to all Gemini Advanced, Business, and Enterprise users; while image generation capabilities powered by Imagen 3 will be accessible to all Gemini users.

Gems are customizable AI-powered assistants designed to perform as experts on topics that align with users' goals. To showcase some of the possibilities that can be unlocked with Gems and get users started immediately, Google has released several premade Gems, aptly named after their specialized tasks. Google's premade Gems include Brainstormer, Learning coach, Career guide, Writing editor, and Coding partner. In addition to chatting with the premade Gems, subscribers to Gemini paid plans can create custom Gems on mobile and desktop in over 150 countries and many languages.

Simultaneously, the company announced that improved image-generation capabilities delivered by Imagen 3 are coming to the Gemini Apps for users across all plans. With the Imagen 3 integration, users can ask Gemini to generate images in several styles, which are watermarked by default with the SynthID technology. Moreover, Google has stated that it has corrected its model's people generation capabilities. Still, the company is being reasonably cautious about a general roll-out, preferring to release an early preview for English-language Gemini Advanced, Business, and Enterprise users. Additionally, there are guardrails in place specific to generations of people, including the refusal to generate photorealistic images of identifiable persons, any depictions of minors, and overly gory, violent or sexual scenes.