While both write and code modes offer the choice to request in-line edits, the split canvas UI is designed with an additional set of shortcuts for those focused on AI-assisted writing and another for programmers. In the demo, Levine showed how the Writer's Shortcut could be used to condense the number of words on a canvas or to try to polish the concept a bit. He also used one of the more light-hearted shortcuts to add some random emoji. On the coder side, ChatGPT can add logs and comments and attempt to troubleshoot issues in a canvas.
ChatGPT saves different versions of the canvas as you iterate, so you can revert to old versions if you ultimately prefer that output. Writers concerned about what they upload is used by OpenAI to train the model should go into their user settings and make sure “model training” is turned off.
By allowing ChatGPT to make changes and suggestions, OpenAI blurs the line between authorship and word management. As someone who works with professional editors every day, I'm skeptical that the Canvas Beta will match their sharp comments and careful guidance. But for people who don't have easy access to human writing partners, I can see how getting synthetic notes on a composition about structure and content would be helpful.
It's worth noting that three people listed as “supportive leadership” on the canvas project are no longer with the company. Former post-training colead and co-founder John Schulman left in August and now works at Anthropic, a rival AI company. Additionally, former Chief Technology Officer Mira Murati and Research Vice President Barret Zoph both stepped down from their roles a week before this launch. At a press event in the OpenAI office following the departure, current Chief Product Officer Kevin Weil reaffirmed the company's commitment to continuing to release software.
“I think 2025 will be the year when agent systems finally become mainstream,” he says. The idea of an AI agent that can not only perform software tasks alongside you, but is also agile enough to be sent into the digital wilderness to do things on your behalf, is simultaneously generative AI's recent past and its anticipated future.
Last year, WIRED covered ChatGPT's plugins that people could use for tasks like booking flights with Expedia or making a reservation with OpenTable — perhaps a step toward more “agentic” AI tools. However, plugins were later discontinued and more limited custom GPT chatbots were launched instead.
With that in mind, the Canvas beta appears to be another attempt to extend AI models with more decision-making capabilities, which could lead to some surprises. During one of WIRED's demos, Levine highlighted a portion of the canvas and requested an edit, and ChatGPT then made an in-line change at the bottom, outside its highlight. “The really interesting thing is often when you highlight a section, an edit is made to that part,” he says. “But ChatGPT has the ability to decide where to edit.”
The closest alternative to OpenAI's canvas tool currently available is probably Google's Gemini integration, which lets you use generative AI in Docs, or Anthropic's Artifacts tool. Chatbots are certainly not dead, but AI companies are now recognizing the limitations of the format and are looking for ways to diversify their software to explore new, sticky user interfaces. Google recently received praise in tech circles for its entertaining AI podcasts; even CEO Sam Altman praised the tool.
With billions of investment dollars continuing to flow through Silicon Valley to AI companies, consumers can expect to see more of these structural experiments that build on existing tools, such as AI podcast hosts and AI document editors, in the coming year. . The chatbot race is far from over, and future iterations of the technology will likely stray far away from that boring chat box, and toward a more versatile approach.
