Originality of AI-Generated digital artworks: A systematic review of the literature

Boy-Guillén L.E.; Sánchez-Palacios L.G.; Cieza-Mostacero S.E. · Aibi, Revista de Investigacion Administracion e Ingenierias (2025) · DOI: 10.15649/2346030x.4582

Resumen

This Systematic Literature Review article aimed to summarize information on the originality of digital artworks created by generative AI, for this purpose, a search was conducted in the following databases: IEEE Xplore, Scopus, ScienceDirect and SpringerLink, finding, after applying inclusion and exclusion criteria, a total of 51 research articles. The main findings were: the most outstanding contributions of generative AI in digital art were efficiency, automation and artistic innovation, with a valuation of 21.88%; the most used techniques for the creation of digital works were Antagonistic Generative Networks and their variants, with a valuation of 32. 14%; the most relevant ethical challenges are related to data bias and discrimination, reaching 32%; the most common training data were tagged images and public bases, with a rating of 32.5%; and finally, the key aspects to evaluate originality in these works were visual quality and diversity, with a rating of 30%. In conclusion, generative AI is key to innovation in digital art, but its use should be managed with caution due to the biases and ethical challenges involved.

Instituciones: Universidad Privada Antenor Orrego