Be a part of our each day and weekly newsletters for the newest updates and unique content material on industry-leading AI protection. Study Extra
Contemporary off the information of its record-setting $6.6 billion funding spherical, OpenAI is updating its signature AI app ChatGPT with a brand new characteristic — Canvas — that permits customers to see, instantly edit, and simply modify simply chosen parts of the chatbot’s outputs in a side-by-side panel view.
The characteristic, constructed atop OpenAI’s GPT-4o mannequin, also can give ideas and implement modifications to the responses within the right-side panel while not having to output a completely new response. It looks like a direct problem to rival AI startup Anthropic’s Artifacts characteristic for its chatbot Claude, which launched in June 2024. It additionally presents a facet panel view to show and run easy Python applications that replace primarily based on the person’s prompts.
Canvas is rolling out to ChatGPT Plus and Groups customers, with subscribers to Enterprise and Edu tiers following in a couple of days. Following the beta, OpenAI plans to make the characteristic out there to all ChatGPT customers.
Customers on the social community X predicted Canvas earlier than its preliminary launch and speculated that OpenAI would quickly launch the characteristic, which turned out to be right.
What ChatGPT Canvas is sweet for
Daniel Levine, product supervisor for Canvas, informed VentureBeat in an interview that generally, the vertical top-to-bottom chat window is simply too restricted for a few of the most typical use instances.
“We know a lot of people use ChatGPT for writing and coding, those are two of the top use cases we see,” Levine mentioned. “But the chat interface is a bit limiting, especially for projects where you want revisions or editing. There’s a lot of back and forth, and comparing changes is hard, so that’s where Canvas steps in.”
Levine is speaking about prompting ChatGPT to edit its responses. With out Canvas, when customers ask ChatGPT to generate a draft e-mail and really feel the response must be longer, funnier, or friendlier, they should immediate ChatGPT once more, and rewrites seem in the identical dialog.
Generally, the draft modifications drastically; generally, the change is delicate, however typically, you need to scroll backwards and forwards to double-check what modified.
OpenAI hopes that with Canvas, the method might be simpler. Customers can, after all, reprompt ChatGPT, but when they solely need a couple of phrases modified, they don’t have to repeat textual content that’s nonetheless very a lot a primary draft to a special doc and edit it themselves.
Tips on how to use Canvas
Canvas will open after customers toggle the mannequin picker, the place they usually select which model of OpenAI’s fashions to make use of. ChatGPT also can detect when Canvas can be applicable to open or when the person sends the immediate,” Use Canvas.” They’ll then immediate ChatGPT to both write one thing or generate code. For instance, if a person needs to put in writing an e-mail to a colleague, they immediate ChatGPT with the request, and it opens the Canvas window with the textual content it simply generated.
Customers can proceed asking ChatGPT to refine the textual content, which is able to replicate on the Canvas display. Individuals also can edit instantly on Canvas, or give directions to the chatbot by highlighting some textual content. On every Canvas window, there’s a set of shortcuts for customers to click on that may alter textual content size, the studying degree, add emojis and even give a last polish.
ChatGPT also can present ideas for the textual content, which is able to present up on Canvas in textual content packing containers, much like how feedback look on Google Docs. It will possibly additionally translate into supported languages.
Canvas appears to be like completely different primarily based on the duty. The writing Canvas appears to be like like a Phrase doc, whereas the coding Canvas contains line numbers for simpler code modifying. Shortcuts for coding can even be completely different. Customers can overview code, repair bugs, add logs and feedback and port to a special coding language via Canvas.
Immediately competing with Anthropic Claude’s Artifacts characteristic
Canvas’s separate window and mannequin collaboration recall to mind one other chatbot with a window that permits individuals to obviously see any modifications made via new prompts: Anthropic’s Claude Artifacts.
VentureBeat’s Michael Nunez reported that Claude Artifacts makes accessible and easy-to-understand interfaces a necessary characteristic in chatbots and referred to as it “this year’s most important AI feature.” In contrast to Canvas, Artifacts is already usually out there to all Claude chat customers.
Artifacts additionally let individuals see how their generated code appears to be like, as they let customers immediate Claude to put in writing code, edit it after which see the fruits of the labor, like a prototype web site or recreation. Canvas simply exhibits customers the generated code and the edits round it.
The brand new interface battleground headed by OpenAI’s Canvas and Anthropic’s Artifacts does level to an issue smaller, third-party AI functions have been making an attempt to resolve, which is how one can make chatbots simpler to learn and use.
Apps similar to Hyperwrite, Jasper and JotBot all generate and supply textual content modifying. Many different software program embrace related modifying options for each code and textual content with the concept that customers don’t want to depart the chat window to make modifications to their work.
Ultimately, options like Canvas and even Artifacts could turn out to be commonplace as extra individuals need to work with chatbots extra streamlined and collaboratively.
“We do think collaborative work is an important part of the workplace,” Levine mentioned. “So we’re taking a first step here.”
