Values, emotions and beliefs within generative AI arts practice patterns in practice: Values, beliefs and emotions within arts practitioners' engagements with machine learning and data mining

Fratczak, Monika; Ochu, Erinma; Medina-Perea, Itzelle; Bates, Jo; Kennedy, Helen

Abstract

Patterns in Practice (PIP) is a qualitative study investigating how practitioners' beliefs, values and emotions shape their interaction and engagement with the use of machine learning (ML) and data mining across three contrasting domains-science, education and the arts. This paper focuses on sharing early insights from exploring culturally situated beliefs, values and emotions of ML and data mining practices within the UK arts sector through a series of qualitative interviews with artists, curators and art commissioners invested in musical, story-based and visual artworks, or combinations thereof, that includes Generative AI systems. Generative AI art has a history stretching back to the 1960s and 1970s when artist pioneers experimented with computer-generated visual art. A computing-driven resurgence emerged in the last two decades with the explosion of available data (accessible often via social media and the internet, in the form of text, images, video and sound), large scale technology investment and increasing computing power, resulting in the emergence of new tools, including AI art generators. The literature indicates that artists use AI as a tool to create artworks and, or, as a topic to critique AI as a concept within artworks. We discuss three emergent narrative themes arising from our empirical analysis of the arts sector, relating to values, beliefs and emotions. This includes a desire to navigate the ethical tensions and limitations of AI tools, such as adopting a 'small data' mindset (over large scale use of data), a desire to improve human-machine collaborations and to minimise the exploitation of minoritised communities and the environment.

• Presentation Conference Type: Conference Paper (unpublished)
• Conference Name: CHI Generative AI Workshop
• Start Date: Apr 28, 2023
• End Date: Apr 28, 2023
• Deposit Date: Oct 14, 2023
• Publicly Available Date: Oct 24, 2023
• Keywords: CCS CONCEPTS; Generative Art; Artificial Intelligence; Values Additional Keywords and Phrases: Machine Learning, Beliefs, Emotions, Cultures of practice
• Public URL: https://uwe-repository.worktribe.com/output/11177044
• Publisher URL: https://generativeaiandhci.github.io/2023