OpenAI has released an update to its image generation tool inside ChatGPT, called ChatGPT Images 2.5, and the most notable addition is a feature called Sketch. Instead of typing a long description of the picture you want, you draw it. Type "@Sketch" in the chat box, doodle something with your mouse or finger, tell ChatGPT what you're going for, and it turns the rough drawing into a finished image.
The company also added a way to leave comments directly on parts of a generated image, so you can point at one detail and ask for a specific change, rather than rewriting your entire request and hoping the rest of the picture stays the same. OpenAI says the new version handles multi step edits better, keeps textures and lighting more realistic, and generates images up to 50 percent faster than the version before it.
Speed is not a minor detail here. Every added second between typing a request and seeing a result is a reason for someone to give up and go back to stock photos or a human designer. Cutting that wait in half means more people will use the tool for actual work, like a first draft of an ad, a mockup of packaging, or a quick visual for an internal presentation.
This launch is also a response to pressure, not just an isolated update. Google has been gaining ground with its own image editing tool, known as Nano Banana, and reports say OpenAI's chief executive told staff internally that the company needed to move faster to keep ChatGPT competitive. That kind of rivalry between two well funded companies tends to benefit everyone else: expect image tools from both sides to keep improving in quick succession over the coming months.
For a business that does not have a design team, this closes a real gap. You no longer need to describe an image in perfect detail to get something usable back; you can sketch the layout yourself and let the tool fill in the rest. That is useful for social media posts, product concepts, presentation slides, or rough packaging ideas, tasks that used to require either a designer or hours of trial and error with text prompts.
It is worth being clear eyed about where this is heading for design work more broadly. Research from the World Economic Forum has already flagged graphic design as one of the fastest declining roles over the next several years, and tools like this are a direct part of why. That does not mean every design job disappears; specialized, high stakes visual work still benefits from a trained eye. But routine, low stakes image production, the kind most businesses need most often, is quickly becoming something anyone can do themselves in a few minutes.