ChatGPT's Canvas now shows tracked changes

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More


ChatGPT’s Canvas feature allows users to edit the chatbot’s responses on the app rather than copying and pasting them to a separate document. 

However, when Canvas launched in early October for its paid tiers, it didn’t let people see what changes GPT-4o made to its responses. OpenAI’s latest update to the feature corrects that. 

The show changes button will show the most recent changes to either the generated text or code on Canvas. It will highlight added information in green and deleted sections in red. 

Tracking changes has always been a good feature of any editing platform; Google Docs and Word documents offer a toggle for users to check what’s been changed. But OpenAI had been planning to roll out updates to Canvas slowly as ChatGPT subscribers get used to it. 

Canvas already offers familiar features like comments, where users can add suggestions or give more instructions for the AI model to follow when editing responses. 

Canvas is still only available on the web version of ChatGPT for ChatGPT Plus, Teams, Enterprise and Edu users. Mac app users and anyone downloading the recently released Windows version of ChatGPT will have to wait until Canvas is rolled out to these standalone apps. 

Currently, people can access Canvas on the regular ChatGPT window rather than in any custom GPTs. 

A much requested feature

OpenAI’s developer X account acknowledged that developer customers have requested a track or show change feature since Canvas launched. 

But while many developers said this was a step in the right direction, Canvas still doesn’t immediately connect to code repositories like GitHub or let users visually see how the edited code works. 

This is one area where ChatGPT competitor Claude from Anthropic and its Artifacts feature excels. Artifacts function much like Canvas; users can begin a prompt on the Claude chat interface. 

When users launch Artifacts, a dedicated window opens where they can manipulate the model’s responses. Artifacts let users replicate websites using the code Claude just generated and edited, so developers can see not only which lines of code have changed but also whether it worked. Artifacts are now available to all Claude users, including those on mobile devices. 

Canvas and Artifacts represent what could be the next phase in the evolution of AI chat platforms and assistants. The Interface War could see other platforms begin to explore how to keep users in the platform instead of opening other dedicated windows for different tasks. 



Source link