×
ChatGPT’s ‘Canvas’ lets you highlight specific sections and get inline feedback
Written by
Published on
Join our daily newsletter for breaking news, product launches and deals, research breakdowns, and other industry-leading AI coverage
Join Now

AI-powered collaborative workspace debuts: OpenAI has introduced Canvas, a new interface for ChatGPT that enhances collaboration on writing and coding projects, marking a significant update to the platform’s visual interface.

Key features and functionality:

  • Canvas displays content in a separate window alongside the AI chat history, allowing users to view working document drafts or programming code while collaborating with the AI assistant.
  • The feature opens automatically when ChatGPT detects a scenario where it could be helpful, or users can trigger it manually by including “use canvas” in their prompt.
  • Canvas includes a menu of shortcuts for common tasks such as finalizing documents, suggesting edits, and adding emojis.
  • Users can adjust writing length and topical complexity using sliders, ranging from short to long and “kindergarten” to “graduate school” level, respectively.

Coding capabilities: Canvas offers several tools to enhance the coding experience and streamline development processes.

  • Coding shortcuts include reviewing code, adding logs for debugging, inserting comments, and fixing bugs.
  • A notable feature is the ability to port code to different programming languages with a few clicks, translating between JavaScript, PHP, TypeScript, Python, C++, and Java.
  • Users can highlight specific sections to direct ChatGPT’s focus, and the AI model can provide inline feedback and suggestions while considering the entire project.
  • The interface allows for easy restoration of previous versions of a working document using a back button.

Technical underpinnings: OpenAI has developed new core behaviors for GPT-4o to support Canvas, enhancing its capabilities and user experience.

  • The research team focused on defining when to trigger a canvas, generating specific content types, making targeted edits, rewriting documents, and providing inline critique.
  • OpenAI trained the model to open a canvas for appropriate tasks while avoiding triggering it for general Q&A scenarios.
  • The team also fine-tuned the model’s editing behavior, teaching it to perform targeted edits when users select specific text and favoring rewrites otherwise.

Availability and rollout: OpenAI is gradually making Canvas accessible to various user groups.

  • ChatGPT Plus and Team users gained access to Canvas globally on the day of the announcement.
  • Enterprise and Education users will receive access the following week.
  • OpenAI plans to make Canvas available to all free ChatGPT users once it exits the beta stage.
  • During the beta phase, Canvas uses a special version of GPT-4o that users can manually select from the model picker in the ChatGPT interface.

Competitive landscape: Canvas enters a market where similar collaborative AI tools are already available.

  • The feature bears similarities to Anthropic’s Artifacts, introduced in June for their Claude AI assistant.
  • Both Canvas and Artifacts serve as scratch pads, visually separating portions of the active working context to prevent them from getting lost in the chat backlog.
  • Canvas’s introduction demonstrates OpenAI’s commitment to enhancing ChatGPT’s capabilities and user experience in response to competition in the AI assistant space.

Future developments: OpenAI has outlined plans for Canvas’s continued evolution and improvement.

  • The company acknowledges that Canvas is still in early beta and plans to enhance its capabilities based on user feedback over time.
  • As the first major update to ChatGPT’s visual interface in two years, Canvas represents a significant step in OpenAI’s efforts to improve the platform’s functionality and user experience.

Implications for AI-assisted work: Canvas’s introduction signals a trend towards more integrated and user-friendly AI collaboration tools.

  • The feature’s ability to seamlessly switch between writing and coding tasks, coupled with its language translation capabilities, could potentially streamline workflows for developers and content creators.
  • As AI assistants become more sophisticated in their ability to understand context and provide targeted assistance, they may increasingly become indispensable tools in various professional fields.
OpenAI’s “Canvas” eases collaborations with ChatGPT

Recent News

Veo 2 vs. Sora: A closer look at Google and OpenAI’s latest AI video tools

Tech companies unveil AI tools capable of generating realistic short videos from text prompts, though length and quality limitations persist as major hurdles.

7 essential ways to use ChatGPT’s new mobile search feature

OpenAI's mobile search upgrade enables business users to access current market data and news through conversational queries, marking a departure from traditional search methods.

FastVideo is an open-source framework that accelerates video diffusion models

New optimization techniques reduce the computing power needed for AI video generation from days to hours, though widespread adoption remains limited by hardware costs.