DALL-E now lets you edit images in ChatGPT

  News, Rassegna Stampa
image_pdfimage_print

OpenAI’s DALL-E is getting a couple of upgrades today that help you shape your AI-generated masterpieces. First, you’ll now find image editing tools when using DALL-E within ChatGPT, both on the web and on mobile. Second, DALL-E now offers preset style suggestions to help inspire image creation — not unlike Android’s AI-generated wallpaper prompts. Because how else could you dream up a sci-fi polar bear on vacation?

DALL-E is on its third generation, and recent updates like this seem to be aimed at making it more user-friendly. DALL-E 3 added integration with ChatGPT, which takes the burden off the user to write a perfectly descriptive image prompt — just tell ChatGPT what you want, and it will do the rest. Today’s update is an extension of that, making it possible for a more casual DALL-E user to tweak and refine their results from within the ChatGPT interface.

The new style prompts in the DALL-E 3 service seem to address a related but slightly different question from a novice user: what the heck do I do with this thing? Sometimes a blank canvas is a little too blank, you know? The new prompts act as a kind of starting point and include styles like woodcut, gothic, synthwave, and hand-drawn, with examples as you mouse over them. Choosing one adds it to your prompt. Then, it’s on you to figure the rest out. So you’re not totally off the hook, creatively speaking.

Recently, OpenAI has also been putting in some work to address a slightly more pressing concern related to AI image creation: trust. Earlier this year, DALL-E 3 started adding visible watermarks and metadata to denote an AI-generated image. A good start, for sure, but both can be easily removed by bad actors. This isn’t only OpenAI’s problem to solve, but it sure is an important one. In the meantime, I hope that sci-fi polar bear has a nice vacation.

https://www.theverge.com/2024/4/3/24120181/openai-dall-e-chat-gpt-image-edit