Google adds collaborative workspace and audio features to its AI assistant, following ChatGPT's lead.
Google has added two major features to its Gemini AI assistant: Canvas, which provides an interactive workspace for document creation, and Audio Overviews, which turns written content into conversational audio formats.
The new Canvas workspace lets users collaborate with AI while creating documents and code. Changes appear in real-time, and users can discuss modifications through an integrated chat interface.
According to Google, Canvas can generate an instant preview of HTML and React code, and users can export finished documents to Google Docs. The feature seems to closely mirror ChatGPT's Canvas feature from last fall.
Converting AI research into AI conversations
The Audio Overview feature transforms written content like documents, presentations, and research reports into podcast-style conversations between two AI hosts who analyze the material. Google demonstrates this by converting output from its updated free Deep Research function into an AI audio podcast.
These audio summaries first appeared on NotebookLM last year. Like other AI-generated content, both the podcasts and the Deep Research feature may contain inaccurate information. That doesn't mean either tool can't be useful, just that you probably shouldn't use them to learn something you don't know because you can't spot mistakes here.
The audio feature currently supports English only, with access via web and mobile platforms. Users can download and share the audio files created. According to Google, both features are now available worldwide for Gemini and Gemini Advanced subscribers.
Google rolls out wave of new Gemini AI features
Google has heavily expanded its Gemini ecosystem since the beginning of the year. Gemini Live adds video and screen sharing support, starting with Samsung Galaxy S24/S25 and Pixel 9 devices, and Google plans to replace Google Assistant with Gemini as Android's default AI assistant, including integration with Samsung apps.
In addition, Google launched three new model variants: Gemini 2.0 Flash serves as the base model, while Flash Lite offers cost-effective processing, and 2.0 Pro provides experimental expanded context handling.
Outside the EU, Switzerland, and UK, Gemini can now access users' Google search history for more contextual responses. The platform has also gained native multimodal capabilities for generating and editing images and understanding YouTube videos.
