• The competition is free to enter for any human.

• Anyone can enter the competition, so long as:

  • The human has commercial rights to the entered work.

  • The image is the result of provided an AI model with a text prompt, optionally with AI-provided algorithmic upscaling.

  • Artwork titles may be chosen by any combination of human-provided and/or AI-provided means.

• Voting will be held in three independent rounds, starting and ending at 00:00 UTC each day:

  • Group 1: Voting open October 24 - October 31

  • Group 2: Voting open October 31 - November 7

  • Group 3: Voting open November 7 - November 14

• Competitors may enter up to three unique works, once per group.

• Competitors may submit entries to any group at any time until voting closes for that group. However, given that we will manually approve entries before they are available for voting, in order to maximize the time that your entry is available for voting you may want to send your submission no later than a day before voting opens for each group.

• Competitors must be at least 18 years of age or have the signed consent of a parent or legal guardian at the time of selection.

• The AI model used for text-to-image generation must be of sufficient complexity, and provided with sufficient inputs, such that the resulting output is not markedly similar to any previously existing image.

• Competitors should retain as much detail as possible of the exact text prompt, model, relevant parameters (i.e. random seed), and upscaling process used. If our judges or advisors have any doubts as to the AI-generated nature of an image, we reserve the right to verify a work’s provenance before selection.

• Individual human competitors may enter once per round; additional entries from the same human will be rejected.

• Each entry will be reviewed by humans prior to approval and display for voting, with a target but not guaranteed response time of under 24 hours. We reserve the right to approve, not approve, or remove any entry for any reason.

• Chosen entries will be notified at their provided email addresses in late November. The human responsible for each chosen submission will have 72 hours to provide written confirmation, or the artwork will forfeit its place in the competition to a runner-up.

• Human-Assisted Art does not claim any ownership rights to submitted images. We will publish all submitted, approved images on our competition site for the purpose of allowing public voting on entries.

• By submitting images to the competition, competitors agree that they have not previously sold any copies of the images, whether digital or physical, and will not do so until judging has been complete. Art featured in the gallery exhibit may not be sold in any additional digital or physical copies in any format.

Additional suggestions:

• Chosen images will be printed at 300 DPI or higher. Keep in mind that at this resolution, 2000 pixels equates to only 6.67 inches, or just under 17 centimeters. Upscale!

  • You may upload a JPG or PNG of up to 60MB to the competition. We recommend getting close to this limit, and/or having a larger version prepared for high-quality printing at 300+ DPI. If your work is chosen, we will work with you on upscaling, sizing, and choice of print materials.

• Given the prevalence of square images as the default format for many AI image models, we encourage the use of non-square aspect ratios. These should be set as input parameters, rather than manually cropping the resulting output.

• The competition organizers may choose matte and framing for displaying all selected works. We will cover these costs and any printing costs in full.

• Do not take the images on this site as guidance or suggestions for your own submission; while these images were AI-generated, they are not intended to represent competition entries.