Dialogues Between Technologists and the Art Worlds

Abstract

Dialogues between technologists and the art worlds Piotr Mirowski, Rida Qadri go/dialogues-art-worlds To be published in a collective position paper on "Arts, International Relations, and Innovations" led by the British Council. Coordinating authors: Hannah Andrews and Aurora Hawcroft.

• This article represents the personal opinions of the authors and does not represent official Google policy.

As generative AI enters the world of creativity, it raises questions and concerns about its impacts on cultural expression and creative practice. While research (including ours) has raised questions on cultural erasure and marginalizations that these tools could perpetuate, AI can also be actively shaped and reconfigured by the communities of its users, allowing for localized adaptation and re-imagination by users, to suit their specific cultural contexts [6].

To help technologists who develop AI, and human-computer interaction researchers who evaluate its impacts on creative artists, and to understand real-world use and impact of technologies, we propose to combine two useful frameworks: the notion of a broader 'art world', and principles of participatory AI. On the one hand, and as sociologist Howard Becker articulates, art exists within an ‘art world’ — a complex network of individuals, institutions, and organizations involved in the production, distribution, and consumption of art [1]. This means that technologies for artistic production will likely impact an entire ecosystem, and not just individual users. On the other hand, participatory AI aims at "incorporating wider publics into the development and deployment of AI systems" [2], with the hope that "participation opens the gateway to an inclusive, equitable, robust, responsible and trustworthy AI."

Examples of participatory studies include studies on Large Language Models (LLMs) for creative writing (e.g., Dramatron) and music production tools (e.g, the Lyria toolbox). In studies on Dramatron, we evaluated the usefulness of LLMs to film and theatre industry professionals including playwrights and screenwriters [5], and through dialogue with writers, we iteratively designed a narratological system for interactive writing and rewriting with LLMs. This study helped us explore how writers would consider the output of LLMs to be derivative, while at the same time express interest in the use of LLMs for literary world building, indicating a gap in the design of creativity support tools — which informed subsequent tool development. In a related study, focus groups with comedians at the Edinburgh Fringe [4] revealed how LLMs could lack cultural value alignment when handling sensitive cultural contexts pertaining to the comedian's identity. Our colleagues at Google Research and Google DeepMind engaged, over extended periods of time, with musicians who use AI tools in their creative process, shaping these tools according to the artists' needs. They evaluated the Lyria toolbox [3] during creative engagements with music artists such as Indian singer and composer Shankar Mahadevan, who tried to fuse his music, inspired by Carnatic, Hindustani, folk and jazz genres, with AI tools, for cross-cultural exploration.

While powerful in revealing ways in which creatives and tech professionals can collaborate for improving design of technology, these studies focused on the artists as central stakeholders for AI technology design [8]. In our recently published work, we expanded that scope to the broader art world. We put visual artists and designers in dialogue with art critics and museum curators specialists of the Persian Gulf [6][7], engaging in critical dialogue about image archives, stereotypes and artistic agency. We closely followed how artists would devise "hacks" and work around existing generative AI tools, then using them in artworks that followed their culturally-relevant visions and lived experience. Following such processes allows us to establish a localised art world with artists and critics, to collect insights into how the reception and generation of technology-enabled art is shaped by histories of creativity, politics and artistic visions, and to situate the discussions on redesigning AI tools for non-Western creativity.

As an outcome of such participatory processes situated within art worlds, we, as technologists, hope to de-centre the focus from just individual users. We hope to see how carefully designed AI tools can both provide artists with new perspectives or approaches to their craft, and adapt AI tools to their creative process, when relevant. The process is as important in art as the output, and knowing about creative processes can also help us identify gaps in design, interface, interactions of generative AI systems and thus present developers with actionable pathways for improving generative AI as a tool for cultural production [7]. We also hope that AI technologies that intervene in artistic expression should be understood in the context of a localized art world [1], and that such cultural specificity will be a crucial element of innovation.

[1] Howard S Becker. 1982. Art worlds. University of California Press: Berkeley, CA.

[2] Birhane, A., Isaac, W., Prabhakaran, V., Diaz, M., Elish, M. C., Gabriel, I., & Mohamed, S. (2022, October). Power to the people? Opportunities and challenges for participatory AI. In Proceedings of the 2nd ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (pp. 1-8).

[3] "Lyria 2: Our latest music generation model", accessed at https://deepmind.google/technologies/lyria/, retrieved on 8 May 2025.

[4] Mirowski, P., Love, J., Mathewson, K., & Mohamed, S. (2024, June). A robot walks into a bar: Can language models serve as creativity support tools for comedy? an evaluation of llms’ humour alignment with comedians. In Proceedings of the 2024 ACM Conference on Fairness, Accountability, and Transparency (pp. 1622-1636).

[5] Mirowski, P., Mathewson, K. W., Pittman, J., & Evans, R. (2023, April). Co-writing screenplays and theatre scripts with language models: Evaluation by industry professionals. In Proceedings of the 2023 CHI conference on human factors in computing systems (pp. 1-34).

[6] Qadri, R., Mirowski, P., & Denton, R. (2025, April). AI and Non-Western Art Worlds: Reimagining Critical AI Futures through Artistic Inquiry and Situated Dialogue. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems (pp. 1-17).

[7] Qadri, R., Mirowski, P., Gabriellan, A., Mehr, F., Gupta, H., Karimi, P., & Denton, R. (2024). Dialogue with the Machine and Dialogue with the Art World: Evaluating Generative AI for Culturally-Situated Creativity. arXiv preprint arXiv:2412.14077.

[8] Qadri, R., Shelby, R., Bennett, C. L., & Denton, E. (2023, June). Ai’s regimes of representation: A community-centered study of text-to-image models in south asia. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency (pp. 506-517).

Authors

Piotr Mirowski, Rida Qadri

Venue

British Council Position Papers