Google has added a Canvas feature to its Gemini AI chatbot that provides users with a real-time collaborative space where writing and coding projects can be refined and other ideas iterated and shared. “Canvas is designed for seamless collaboration with Gemini,” according to Gemini Product Director Dave Citron, who notes that Canvas makes it “an even more effective collaborator” in helping bring ideas to life. The move marks a trend whereby AI companies are trying to turn chatbot platforms into turnkey productivity suites. Google is launching a limited release of Gemini Live Video in addition to bringing its Audio Overview feature of NotebookLM to Gemini.

“What started as a chatbot has now taken over Assistant’s role,” Android Police says of Gemini, noting that it has “permeated into Google Workspace, and also rivals other Search tools like Lens and Circle to Search.” The new live video AI features take that development further, making it possible to share multimodal screensharing capabilities live with Gemini AI.

Google’s initial Gemini Live Video release “will focus on Gemini Advanced subscribers with a Google One AI Premium plan,” Android Police says, adding the hope that “limited demos will be available for free-tier users to drum up excitement for the subscription.”

Its release coincides with “glimpses” of Google’s Project Astra in action. Announced in May 2024, Project Astra “represented the company’s short-term vision for a real-time assistant capable of multimodal audio and video prompting,” Android Police writes.

In many ways, Gemini Live is the practical application of Astra in the real world, allowing users to “‘see screens and camera feeds in real-time for some Google One AI Premium subscribers now,” The Verge writes.

While Android Police notes that for some users “neither Share screen with Live nor Astra were available,” it adds that scattered early glimpses “suggest we are witnessing the early signs of a staggered worldwide release through server-side changes, provided Google keeps its word and follows through.”

As for Canvas, TechCrunch writes that it is “similar in concept to OpenAI’s identically named Canvas tool for ChatGPT and Anthropic’s Artifacts.” Dedicated workspaces such as these “can offer more precision than text-based interfaces alone, as well as provide a way to preview code” and other projects in real time.

Canvas does skew heavily to coding, however, with programming-centric features like an HTML preview generator, built-in React code library for JavaScript and other Web app prototypes. “Users can ask Gemini to make changes to a preview, and Canvas will iteratively refresh it,” TechCrunch notes.

Gemini already simplified programming and software development, including code generation, debugging and code explanation, Citron writes in his blog post, explaining that “now, Canvas streamlines the process of transforming your coding ideas into working prototypes for Web apps, Python scripts, games, simulations and other interactive apps.”

TechRadar reports that Google has also added Gemini-powered “contextual smart replies” to Gmail for business subscribers, “allowing its Gemini AI tool to do the heavy lifting (and thinking) when it comes to writing better emails.”

In addition, “Google’s Gemini app now lets you generate Audio Overviews based on Deep Research,” writes The Verge. “That means you can turn the in-depth reports generated by Gemini into a conversational podcast featuring two AI ‘hosts.’”

Related:
Google Gemini Introduces Collaborative Canvas and Podcast-Like Audio Overviews, SiliconANGLE, 3/18/25
With Google’s Gemini, Create Podcast-Style Discussions from Your Docs, TechRepublic, 3/21/25
Google’s Gemini Deep Research Is Now Free — I’ve Got 7 Quick Tips to Start Using It Now, Tom’s Guide, 3/21/25
Is Gemini Canvas Better Than ChatGPT Canvas?, TechRadar, 3/20/25
Forget ChatGPT Canvas — I Just Tried Gemini Canvas and I’m Floored By the Difference, Tom’s Guide, 3/19/25