2022 Duke AI for Art Competition

Using algorithms to generate art promotes deeper understanding of the ways that artificial intelligence (AI) is changing the visual and media-based world around us and provokes inquiry into the endeavor and humanity behind creativity. In this spirit, the Duke Center for Computational Thinking held the 2022 Duke AI for Art Competition.

All Duke University students, faculty and staff were eligible to submit artworks generated by artificial intelligence methods, systems, networks and algorithms. Individuals or teams from any discipline or field could participate, regardless of prior experience.

Submission details

While submissions were encouraged from all varieties of artistic media, the final work had to be submitted in the form of one or more digital images. Submissions were digital image files in PNG format with a resolution of 4096 pixels in the largest dimension. The artwork had to be entirely generated by artificial intelligence software running on a computer. Submissions required an accompanying statement outlining both the conceptual foundation of the work and the specific technical methods employed in the creative process. The statements had to cite the source of any art used to seed the entry and reference the source of any software that was not created directly by the submitting individual/team. Submissions had to include the software that was used to generate it, or a link to the software in an online format. The generation of the art had to be reproducible by running the submitted software. Students were encouraged to submit works that have been previously developed in courses or projects; the description of the development had to include acknowledgement any collaborators.

First, second and third place prizes were awarded. The submissions were judged by Duke Visual Art Faculty, Research Computing staff and external AI researchers. Competition winners presented their artworks at the Center for Computational Thinking Symposium on Thursday, March 3, 2022. Each award winner gave a short talk on the process, methods and final outcomes of the artwork.

Opportunities to learn about the methods of art generated by artificial intelligence methods, systems, networks and algorithms were held on the CCT's YouTube channel in advance of the competition (see below).

Questions? Professor Augustus Wendell at augustus.wendell@duke.edu.

Learning Opportunities

Duplicates, Iterations, Data and Visuality

Virtual learning experience (VLE) with faculty in conversation exploring the shared domains of visual art, generative creative algorithms and artificial intelligence systems.

Innovation, Influence and Originality: Artificial Intelligence in the Creation of Visual Art

A more technical learning experience at which participants were taught how existing AI methods may be used to generate visual artworks.