Complex algorithmic, data-driven infrastructures have rapidly inscribed themselves into almost all processes of everyday life. And although they remain hidden underneath the perceptible surface, they shape the way we perceive the world around us, but also ourselves, how we think and act under the digital condition (Stalder, 2018). Art is not unaffected by this either. The application of AI in creative, artistic processes is becoming increasingly popular, which at the latest became obvious with the introduction of Dall-E 2 or stable diffusion in the spring and summer of 2022. This raises the question of how artists deal with complex algorithmic and data structures in such articulative processes and to what extent new forms of the subjection result from this. As art always operates under prevailing socio-technical as well as socio-cultural conditions, it reflects on social values and norms (McLuhan, 1964). For example, art also deals with diversity issues and discrimination problems that are reproduced or even reinforced by digital technologies (Bajohr, 2022). This bias is caused by the underlying training data with which the models were fed. „They are central to how AI systems recognize and interpret the world. These datasets shape the epistemic boundaries governing how AI systems operate, and thus are an essential part of understanding socially significant questions about AI“ (Crawford & Paglen, 2019). The paper demonstrates in three steps to what extent the role of "AI" in articulative processes can be understood as explorative in relation to diversity issues and discrimination problems: In the first step, the relation between machine learning and big data is discussed. Following on from this, the second step highlights the extent to which this data can be seen as biased (Stark et al., 2021; Crawford & Paglen, 2019). In the third step, empirical material in the form of qualitative interviews and ethnographic observations is used to show how the artistic approach to such complex infrastructures and technologies is shaped and what the role of data is in this process. The resulting works not only make the inscribed complexity visible, but can also be experienced beyond it. They can contribute to important discussions about AI, diversity, and discrimination in the public sphere and sensitize the audience to these issues. In turn, from a media-educational point of view, these artistic practices may provide access to the complexity of data-driven algorithmic systems (Verständig & Ahlborn, 2020).

Bajohr, H. (2022). Malen nach 0 und 1. REPUBLIK. https://www.republik.ch/2022/05/07/malen-nach-0-und-1 (last access: 14.12.2022) Crawford, K. & Paglen, T. (2019). “Excavating AI: The Politics of Training Sets for Machine Learning (September 19, 2019) https://excavating.ai (last access: 14.12.2022) McLuhan, M. (1964). Understanding Media: The Extensions of Man. Mentor. Stalder, F. (2018). The digital condition (V. Pakis, Trans.). Polity. Stark, L., Greene, D., & Hoffmann, A. L. (2021). Critical Perspectives on Governance Mechanisms for AI/ML Systems. In J. Roberge & M. Castelle (eds.), The Cultural Life of Machine Learning (p. 257–280). Springer International Publishing. https://doi.org/10.1007/978-3-030-56286-1_9 Verständig, D., & Ahlborn, J. (2020). Decoding Subjects? Über Subjektivierung und Kreativität im algorithmischen Zeitalter. In J. Holze, D. Verständig, & R. Biermann (eds.), Medienbildung zwischen Subjektivität und Kollektivität (Vol. 45, p. 77–94). Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-31248-0_5