Towards a Principled Evaluation of Likeability for Machine-Generated Art

Lia Coleman, Panos Achlioptas, Mohamed Elhoseiny

Research output: Chapter in Book/Report/Conference proceedingConference contribution


Creativity is a cornerstone of human intelligence and perhaps its most complex aspect. Thus, it is very interesting to understand how AI is already being used by professionals in creative domains like the arts and fashion. Namely, do artists actually like AI-generated “paintings"? In this study we collect and analyze responses on these questions from various contemporary artists and compare them to more naive, crowd-sourced ones. We highlight the importance of considering artists’ opinion when evaluating AI-based art, and present a promising approach for researchers to do this easily.
Original languageEnglish (US)
Title of host publication32nd Conference on Neural Information Processing Systems (NeurIPS 2019), Montréal, Canada
StatePublished - 2019

Bibliographical note

KAUST Repository Item: Exported on 2020-10-01
Acknowledgements: We want to thank the artists who helped us with our research: Brooke Cheng, Dwayne Jones, Joseph Wilk, Luisa Fabrizi, Mark Hernandez, Taís Mauk, Michelle Cheung, Julia Peter, Francisco Rojo, Mathilde Mouw-Rao, Iain Nash, and Achim Koh.


Dive into the research topics of 'Towards a Principled Evaluation of Likeability for Machine-Generated Art'. Together they form a unique fingerprint.

Cite this