We already know that AI art generators are not all created equal. Many of the free AI art generators produce bizarre blurry results that look like an accidental image shot on one of the first camera phones while more powerful tools can create convincing photorealistic images and recreate artistic styles. Now a creative who uses the tools has made a direct comparison of three of the best AI art generators available, and the results are fascinating.

The experiment shows very different results from each tool, offering an insight into how each one interprets prompts. If we didn’t know better, we’d almost think each AI art generator has its own personality and style like a human artist. What’s clear is that each tool has its strengths and weaknesses, and sometimes even a recognisable look, based on its algorithms and the database of images it’s been fed, and this is something that artists may be able to take advantage of (see our guide to how to use DALL-E 2 if you want some tips).

Berlin-based Fabian Stelzer (opens in new tab), who describes himself on Twitter as a ‘prompt intern’ working on three AI-based projects, carried out the image comparison experiment using the text-to-image AI art generators Midjourney, DALL-E 2 and Stable Diffusion. He entered the same prompts on each tool and used a 1:1 aspect ratio for the resulting images. 

With prompts ranging from “low poly game asset, Cthulhu monster, 2000 video game, isometric view” to “1990s clip art of a laughing crazy fax machine, windows 3.1, MS-DOS, early computer clip art”, the results that Stelzer shared in his Twitter thread allow us to compare how the tools handle different types of requests.

Midjourney’s creations often feel very dark – almost apocalyptic. After all, this is the tool that was used to create the “last selfie on Earth” images that were going around recently (see our roundup of the weirdest AI art). We think this AI art generator definitely needs counselling, but it also seems to often produce the most natural results when it comes to artistic styles, particularly with textural details. Any artefacts appear natural, whereas in DALL-E 2 artefacts often look like obviously digital glitches.

DALL-E 2 has a tendency to throw in random invented words, but it seems to be the best tool for creating photorealistic images and for handling facial expressions. Meanwhile, Stable Diffusion seems to often produce the cleanest results – Stelzer notes that it can create incredible photos too but that you need to be careful not to “overload” the scene. It’s also good at recreating the artistic styles of specific artists.

Stelzer says he sees these and other AI art generators much like musical instruments, each with its own range and timbre. Midjourney is like an analogue Moog – beautiful but with a limited range while DALL–E 2 has a huge range but a more obviously digital result “These image synths are like instruments – it’s amazing. We’ll get so many of them, each with a unique ‘sound’,” Stelzer says, adding: “You want to play’n’prompt DALL-E / Midjourney / StableDiffusion individually to their own strengths.”

The three tools have other differences too, beyond the initial results of how they handle prompts. DALL-E 2 offers a powerful inpainting feature that allows you to edit part of an image, while Midjourney has a big, active community of users for support and inspiration.

Stelzer believes that AI art generators will revolutionise creative work in ways we haven’t seen since the advent of photography – “what photography was to painting, image synths are to photographs,” he says, and predicts that soon anyone will be able to create film-like content by typing it out. He’s using AI to create community-narrated ’70s sci-fi film SALT (opens in new tab) and the experimental Twitter game @battleprompts (opens in new tab), in which players summon monsters through prompts – it’s a fascinating glimpse at the creative possibilities that could emerge through AI-generated art and AI-generated video too.

Read more: