OpenAI Dall-E 3 Review: Generative AI for Fanciful and Fun Illustrations

 With Dall-E, OpenAI paved the way for generative AI, which converts a text query into an image. Although there is more competition now, version 3 of the service remains viable.

In my experiments comparing it to Adobe Firefly and Google ImageFX, I discovered that Dall-E 3 consistently produced the most realistic and fascinating photos, as well as the most strange fantasies. It's tricky, but it's most likely to produce decent, useable results on the first try, especially if you're seeking for enjoyable AI hallucinations rather than failures.

Dall-E was also the best at encouraging you to be wacky and see what is possible. I'm confident that there are designers, artists, programmers, and others who can providetheir visions to fruition, but I'm not that skilled. So for me, Dall-E is a marvel.

Dall-E fosters over-the-top prompt engineering, in which individuals submit lines of prose that are somewhere between a vignette and a short story -- prompts that some rivals reject as being too long. Check out this collective vision of Kansas settlers dreaming of a prosperous future after conquering nature and Native Americans. This is an image created from a 186-word prompt. It's an intriguing form of computer-assisted creativity, and Dall-E is the greatest tool for the job that I've tested.


Dall-E 3 is only available as part of the premium ChatGPT Plus service, which costs $20 per month and includes access to a more responsive ChatGPT chatbot as well as OpenAI's helpful GPT Store, which contains custom versions of its AI products.

OpenAI states that it may use content given to Dall-E 3 to improve the model's performance, that it distributes content with a limited number of "trusted service providers," and that it does not sell data or share content with third parties for marketing purposes. You can also file a privacy request asking OpenAI to stop training on your data or cancel your account. For further information, see OpenAI's general privacy FAQ and the main privacy policy.

How Safe Haven tests AI image generators

Safe Haven takes a practical approach to evaluating AI image producers. Our goal is to establish how well it compares to the competition and which purposes it best serves. To accomplish this, we give the AI cues based on real-world use scenarios, such as rendering in a specific style, integrating pieces into a single image, and managing longer descriptions. We rate the picture generators on a 10-point scale based on characteristics such as how well images match prompts, creativity of output, and reaction time. See how we test AI for more.

How good are the images, and how well do they match prompts?

ChatGPT outperforms the other text-to-image AI technologies I tried in terms of providing useful, interesting, and credible outcomes. It still makes lots of mistakes, such as a pickleball player whose paddle sprouts out of his skull instead of the paddle grip, but the results compelled me to investigate further rather than closing the browser tab. It performed better in dynamic settings, with contact and exchanges between participants, and with mood.

ChatGPT is an essential component of Dall-E. It intensifies your prompts and adds florid verbiage to make the outcomes more dramatic. It also allows for a more conversational usage style: you can request a picture and then request a change without having to resubmit your entire query.

That allows Dall-E 3 to surpass competitors such as Adobe's Firefly and Google's ImageFX when it comes to converting your input into what you want and accurately combining various pieces. For example, Dall-E 3 was the only AI image generator I examined that produced a dragon soaring over a castle, spouting fire and gripping a white fluffy sheep in its talons. Granted, it was holding the sheep lovingly, possibly in reaction to OpenAI's anti-violence policies, but it was close.



Do not anticipate perfection. Dall-E missed many details. In an image of a dog walker overwhelmed by too many dogs, the human is delightfully outmatched by his hundreds of canine companions. However, if you look closely, you can see common AI issues: one of the dogs had two heads, another was a cat, and others had strange difficulties with their legs, ears, and tongues. But the image remains intriguing.

Another issue I encountered was the difficulty to generate lifelike visuals. Dall-E 3 tends to generate dreamy settings, frequently in an artistic style rather than a photorealistic one. Personally, I'm cool with it, because photorealism frequently exposes the most cringe-worthy flaws in AI photography. However, photorealism is an important use case, and I had
It was difficult to direct Dall-E toward it.

I could never capture a good shot of a British Navy sea captain on a sailing ship. Dall-E couldn't figure out how people used telescopes, had no idea what a sextant was, and once populated a ship with 12 captains and no crew. The gist of what I needed was often present, but it was not convincingly presented.

Dall-E 3, like its competitors, cannot count. When I asked for a single pool ball, it showed me an 8 ball that was half green and half white. That is not something you would see in the neighborhood billiards hall.

ChatGPT Plus customers additionally get access to at least ten bespoke logo generating GPTs based on Dall-E that are optimized for that task and accessible in the GPT Store.

Images were repeatedly ruined by such detailed defects. Nonetheless, I continued to like the results. Exploding balloons, biblically accurate angels, red-tailed hawks, coffee shop logos—nothing was ever flawless, but it was usually enough.

How engaging are the images?

Very engaging. Dall-E 3 repeatedly provided visually appealing photos. Even when there were challenges, I usually appreciated them. They made me laugh at times and focus on details.


However, Dall-E 3's maximalist language style might be objectionable at times. When asked for an image of a doctor and patient surrounded by medical equipment, a dozen monitors recorded cardiac and respiratory data. One of the PCs' keyboards featured approximately 100 keys.

People who are very emotional can also appear insane. My request to render a frustrated person standing behind a box of cleaning goods resulted in two people who were more outraged than frustrated, and one who was downright demonic.

Conclusion

Dall-E 3 is an impressive tool that can inject some creative fun into your life and do useful image creation work. As with all text-to-image generation tools, it's prone to errors, but Dall-E 3 offers the best results among rivals in my testing. You'll have to decide for yourself if the relative quality -- and the best version of the ChatGPT chatbot -- is worth $20 a month of your budget.