How You Can Try the New ChatGPT Image Generator

DALL·E 2 introduced many of us to powerful image generation using artificial intelligence. Once OpenAI lifted the waitlist restriction, anyone could type a request into a text field and see their ideas brought to life using artificial intelligence. Now OpenAI will soon introduce DALL·E 3 to the world , not only improving DALL·E 2’s image generation but also directly integrating with ChatGPT.

What’s new in DALL E 3

OpenAI didn’t go into too much detail with the DALL·E 3 announcement, but we do know there are some significant improvements over DALL·E 2 . First, OpenAI claims that DALL·E 3 is much better at handling all aspects of any given clue. Where DALL·E 2 might discard one or two concepts during processing, DALL·E 3 supposedly produces content that accommodates all requests.

OpenAI also touts overall improvements to DALL·E 3’s results. Given the same query, DALL·E 3 should produce more accurate, brighter, and more impressive results than DALL·E 2.

But perhaps most notably, DALL·E 3 runs on ChatGPT Plus. If you pay for the premium version of ChatGPT, you can create images directly in your chats. In addition, ChatGPT will enhance your suggestions, offering more detailed queries that will allow DALL·E 3 to create a better image. If you’re not happy with the end result, ask ChatGPT to customize it to your liking.

OpenAI demonstrated the new ChatGPT/DALL·E integration to The Verge’s Emilie David . In one scenario, they asked ChatGPT to create a logo for a ramen restaurant in the mountains. ChatGPT took on the challenge of coming up with a longer clue, then used it to generate four variations using DALL·E 3. David’s favorite was “the image of a mountain with ramen snow caps, broth flowing down a waterfall, and pickled eggs on the ground like garden rocks “, although they admitted that the result was more suitable for a consumer product than a restaurant.

DALL·E 3 pays special attention to safety

OpenAI says DALL·E 3 is designed to reject requests to create a public figure, at least by name. They also take the time to prevent the emergence of visual biases, propaganda and misinformation.

The company is also looking at ways to flag when an image was created using AI, starting with a “provenance classifier” that will help them internally identify when DALL·E 3 was used to create an image.

In addition, OpenAI will allow artists to block the use of their work in future versions of these image generation models by submitting a work to which they own the rights. Future versions of DALL·E will not produce results that resemble the locked image and style. This is a good development as the world considers the ethics of creating AI art .

When can I try DALL·E 3?

DALL·E 3 has not yet been released as it is still in the “pre-research” stage. However, OpenAI says the service will be available to ChatGPT Plus and Enterprise customers starting in October.