Image created by Peter von Stackelberg using Midjourney

Prompt-driven artificial intelligence (AI) art generators use text, images, or other types of inputs to initiate the generation of an image.

The most current AI art generators are predominantly text-based systems. However, with the rapid evolution of this technology, it is becoming more common for AI art generators to use a combination of text and images.

ApplicationCostEase of UseURL
DALL-E 2SubscriptionModerate
Dream by WomboFree/SubscriptionEasy
Stable DiffusionFree/SubscriptionModerate
This list was compiled on March 16, 2023 and, given the rapid evolution of AI art generators, will likely change.

A text prompt for an application like Midjourney should be a description of what the user wants to image to include. This can include:

  • subject matter such as landscape, portrait, or other image
  • colors used in the image
  • objects that appear in the image
  • the visual style of the image, such as photorealistic, comic book style, oil painting, or other style

The AI system uses the information in the prompt to generate an image, while incorporating its own interpretation of the prompt. These four images were generated in Midjourney using “ruins of an alien temple built into sandstone cliffs surrounding a deep valley with sand dunes” as a prompt.

The four preliminary images generated by Midjourey from the prompt “ruins of an alien temple built into sandstone cliffs surrounding a deep valley with sand dunes”

The four preliminary images created using identical prompt, but with the style of the artwork added, were black-and-white images with more of a graphic novel appearance.

Images generated using Midjourney from the prompt “ruins of an alien temple built into sandstone cliffs surrounding a deep valley with sand dunes, black-and-white line art, graphic novel style
The final version of a landscape with the ruins of a temple in a desert valley, created by Peter von Stackelberg using Midjourney.
The prompt was modified to change the image to an anime style and the aspect ratio (width to height) to make it wider.
Images created by Peter von Stackelberg with Midjourney using the prompt “street scene of a futuristic city at night, tall buildings, cafes and shops with neon signs, streetlights, anime background, style of tekkonkinkreet –stylize 500 –version 4 –ar 16:9”

A totally different landscape, in this case a futuristic city lit by neon signs and streetlights, is rendered in a similar anime style to the previous desert landscapes. Having the AI application scale up the image selected by the user — in this case, the image on the lower left — can quickly produce a final image.

Image created by Peter von Stackelberg using Midjourney.

NOTE: The author of this article, the Prunty-Russo Media Lab, and Alfred University do not endorse or recommend any of the companies or AI art generator applications mentioned in this post, nor do the author, Media Lab, or university receive any form of compensation for including or excluding any company or application in this or any other post. The companies and/or sites are identified for informational purposes only. AI technology is evolving very rapidly, so information about AI applications is likely to change quickly. Users are entirely responsible for determining the validity of any information and for their use of any application or website.

Leave a comment

Your email address will not be published. Required fields are marked *