Attuning to computational creativity: A socio-technical imaginary
Pfefferkorn, J. and E. K. Sunde, ∀ISION_E: Drawing a Vision for Every Extended Intelligence (2025).
- Human-AI Co-creation
- Computational Creativity
- Computational Aesthetics
- Generative AI
- World Models
Abstract
Louise Amoore and colleagues (2024) recently wrote that the generativity of GenAI systems extends beyond the creation of a ‘novel’ output, and into the creation of a new logic/ politics of AI. This paper proposes to build on this theoretical foundation to illustrate how GenAI systems forge an emerging aesthetics that arises from human-machine interaction.
Our positioning of aesthetics is largely phenomenological – as both sensing and sense-making (Fuller and Weizman, 2021). We put forward ‘attuning’ as a necessary mode of perceiving oneself as being in creative partnership with AI tools. Building on Anna Munster’s recent work DeepAesthetics (2025), we argue that the necessity of attuning stems from the recognition of GenAI systems as not only technological, but also culturally constructed as well as culturally constructing. This socio-technical ensemble exceeds human-centred imaginaries. We argue for a recognition of the creative agency of AI models, alongside their material specificity.
Through a discourse analysis of publicly available material from OpenAI on Sora and Sora 2 Pro, we identify three key terms: seamlessness, consistency, and simulation. We find that each term is represented by OpenAI as a ‘neutral’ technological affordance. Our analysis results in a counter perspective, finding that these processes are embedded with human-centred imaginaries from development through to interaction and content circulation. To address this disjunction, we put forward ‘attunement’ as mode of engagement. This enables users to perceive the entangled social, cultural, and technical layers that inform the aesthetics of AI and, consequently, building an understanding of the creative capacity of GenAI. Our paper makes two propositions. First, we consider creative agency through the user-model interaction and questioning the parameters of creative partnership. Second, it positions the generativity of models like Sora not as output, but as emerging aesthetic processes.