Generative ominous dataset: testing the current public perception of generative art

Veiga, Pedro Alves

http://hdl.handle.net/10400.2/15751

Use this identifier to reference this record.

Name:	Description:	Size:	Format:
3623462.3623475.pdf		6.08 MB	Adobe PDF	Download

Send Feedback

Authors

Veiga, Pedro Alves

Abstract(s)

The advent of generative AI artworks has paved the way for groundbreaking explorations in the realm of digital creativity. This article delves into the multifaceted dimensions of G.O.D., an abbreviation for the art project Generative Ominous Dataset. G.O.D. aims at critically engaging with contemporary AI generative image systems and their intricate interplay with copyright issues, artistic autonomy, and the ethical implications of data collection, unravelling its conceptual underpinnings and its implications for the broader discourse on artificial intelligence, artistic agency, and the evolving contours of digital art. G.O.D. is a generative artwork, entirely coded in Processing, and developed within a/r/cography, a creative research methodology. G.O.D. scrutinizes and questions the ethics of contemporary text-to-image AI-based systems, such as Midjourney, DALL-E, or Firefly. These systems have been at the centre of controversies concerning the datasets used for their training, which encompass online sourced copyrighted materials, without authorization or attribution, masking questionable approaches with technological dazzlement. Many artists and authors find their works repurposed by these systems for the mass production of digital derivatives. G.O.D. aims at critically exposing art audiences to these concerns.

Keywords

Generative art Dataset Ethics Copyright

URI

http://hdl.handle.net/10400.2/15751

Citation

Veiga, P.A. (2023). Generative Ominous Dataset: Testing the Current Public Perception of Generative Art. In Proceedings of KUI '23, the 20th International Conference on Culture and Computer Science: Code and Materiality, Lisbon, 1-10.