Abstract
Artificial intelligence, or AI for short, is one of the most significant developments of the 21st century. Even though this book was published in German, the topic of AI and the efforts of cultural institutions to carefully integrate AI into their work are worthy of being noticed in non-German-speaking countries as well. The editors, Martin Lätzel, Director of the Schleswig-Holstein State Library, and Tobias Hochscherf, Professor of Audiovisual Media at Kiel University of Applied Sciences, open their anthology with this indisputable statement. Together with 14 other authors, they aim to shed light on the “ubiquitous, dynamic and accelerating” (p. 10) challenge facing society as a whole, as well as highlighting the dormant potential and opportunities for art and cultural production and its mediation with AI. Above all, however, they want to achieve one thing: take away the fear of dealing with AI.
The editors are guided by a central thesis, “AI can help in the transformation process of culture” (p. 101). This assumption was also the basis of the research project to which the volume owes its creation and which Kiel University of Applied Sciences and the Schleswig-Holstein State Library carried out from 2019 to 2020 under the title AI In Cultural Infrastructure Institutions (in short: AI & Culture). The editors collaborated with numerous cultural institutions, including the European Hanseatic Museum in Lübeck, the Molfsee Open-Air Museum, the Kiel City Archive, the LINK Network in Lower Saxony, and the Schleswig-Holstein State Museums Foundation. This anthology brings together a total of 14 essays and reflections, workshop reports, speeches, and documentaries. This (intentional) mix of practical reports and theoretical exploration is one reason why the quality of the contributions varies quite a lot.
But what exactly is AI? The first part of the volume (Theory: Requirements) attempts to answer this question in four texts that define the framework conditions for the implementation of AI in the cultural sector. The first contribution by business information scientists Stephan Schneider and Irina Loza stands out in particular as it is the only chapter that does not originate from the field of cultural studies. Schneider and Loza define AI as “systems (…) that perceive their environment rationally, i.e., process data based on rules and initiate measures based on this in order to achieve predefined goals” (p. 20). Schneider and Loza distinguish between four different types of AI, which they characterize as “weak AI” (1, reactive actions, 2, limited memory storage) and “strong AI” (3, capable of social interaction, 4, capable of self-awareness) (p. 21). While humans have all these forms of intelligence, there is still no AI that can act empathically or with foresight. It is still weak AI that humans use to their advantage and for which they themselves set the meta-rules.
Matthias Bauer, Professor of Media Cultures at the University of Flensburg, relativizes this statement insofar as he points out results in his critical review of the two books The Creativity Code by Marcus du Sautoy (Du Sautoy, 2021) and Theory Of The Digital Society by Armin Nassehi (Nassehi, 2019), which outline the machine's path to a strong AI: In the Sony computer lab in Paris, 20 robots are already interacting with their mirror images and inventing new words in the process (p. 75). Bauer's contribution raises important questions with regard to the framework conditions of “strong” (p. 73) intelligence. He confronts technological progress with “cultural codes” (p. 73), which attribute meaning to actions in the first place—in other words, with the set of meta-rules that make algorithmic work “meaningful” (p. 73) in the first place. His well-founded contribution remains in the sphere of theoretical reflection and, apart from a few art-philosophical comparisons, has few references to practical cultural work.
The question of the role of cultural policy in these developments is then addressed in two cultural policy articles. These round off the first part of the volume—with quite different approaches. Dirk Schrödter, Head of the State Chancellery of Schleswig-Holstein and Minister for Digitization, contributes witty impressions of ChatGPT and is convinced of the indispensability of digitization in the cultural sector, while Henning Mohr, at the time Head of the Institute for Cultural Policy at the Cultural Policy Society, emphasizes the necessity of fundamental transformations. The latter states, “There is much to suggest that the principle of digitality in the cultural sector has not yet been understood at all” (p. 38). Mohr's plea is aimed at a change in thinking that regards AI not just as an optional add-on, but as a paradigm shift. He links this to an understanding of culture that sees culture as a renewing social force and less as a custodian of “bygone days” (p. 39). The question of whether this understanding can actually represent the entire cultural sector immediately arises. In this volume, at least, the majority of the contributors come from institutions that manage and deal with exactly that heritage of “bygone days” (p. 39), or more precisely: archives and museums.
In its second part (Application Research), the volume presents field reports from employees and partners of the AI & Culture project. Tobias Hochscherf opens with a description of this project and its (political) premises. He outlines four fields of action for AI in the cultural sector, which also serve as a guide for the other contributions in the second part: marketing, quality improvement, process optimization, and art as well as cultural production (103ff.). Irina Lozas and Tanja Timmler's report on a data collection project at the European Hanseatic Museum Lübeck and the Molfsee Open-Air Museum shows how AI can potentially be used to more accurately determine visitor motives; Dennis Przytarski documents a pilot project at the Kiel City Archive and provides insights into attempts to semantically index the collection using AI software; Annika Hartmann describes several fields of AI use, namely marketing, big data and creative AI. Unfortunately, these three chapters do not have great news value: They are snapshots that leave questions unanswered and the gain in knowledge is rather limited.
The contribution that opens the third part of the volume (“Results: Potentials”) is beneficial for cultural managers. Entitled, AI As A Task For Cultural Management, Martin Lätzel makes suggestions for AI-supported audience development and draws an interesting comparison: just as cultural managers in the 1990s had to define their self-image in the tension between economics and culture, they now have to find strategies in the digital 21st century to mediate between data protection and effective marketing campaigns with AI. What is interesting about Lätzel's approach is that he sees the future of cultural management in the “curation” (p. 152) of this data, which also means that the work is becoming more technical. Karen van den Berg (2007) had already pointed out the role of cultural managers as mediators and translators in the field of tension between art and business. Lätzel places this curating in a digital context and provides an approach to thinking about the profession further. However, this also raises the question of how cultural managers can cope with this challenge if they are not IT professionals.
The subsequent three chapters are (mostly) short project documentations that deal with various fields of action in the education and cultural sectors. They have in common that they see a lack of knowledge and a shortage of resources as important reasons for the cultural sector's hesitant approach to AI: Christian Möller reports on experiences with persona design in (digital) cultural marketing, Ulrike Aumüller, Moritz Larsen and Doris Weßels evaluate the use of AI in (student) text production, Tabea Golgath sheds light on (digital) networking with the help of AI in art and culture. The format and style of these contributions (project and workshop reports) are similar to the texts in the second part of the volume. It is not clear what prompted the editors to arrange the contributions in two different sections in their volume.
As editors, Lätzel and Hochscherf show an awareness of the problem: “Due to the rapid technical development, some specific examples in this volume are already partially outdated by the time it is published” (p. 185). Attentive readers will notice this too. They will wonder why the editors have apparently included some outdated chapters in their volume. At the same time, they formulate a memorable challenge: the cultural sector should contribute to “promoting the acceptance of new technologies and alleviating society's concerns about the impact of technology on social and cultural aspects” (p. 187). What does this mean? Should the cultural sector contribute to the social acceptance of technologies developed by global IT companies that are at home in “economic exploitation logics” (p. 47)? Is this really its task? Would it not be more appropriate to situate (critical) advocacy, the discourse of pros and cons about an all-pervasive technology, in the cultural sector? This seems all the more reasonable given that, as the editors write in their introduction, “for the first time, humans no longer understand their tools to any great extent” (p. 9). If that is so, it would make sense to take a more critical view of digitalization and the AI boom. But that is not what the volume does.
This leaves the impression of qualitative heterogeneity; not all contributions prove to be equally insightful for work in the cultural field. This may also be due to the fact that the topic of AI can basically no longer be mastered without the involvement of IT experts—not even in cultural studies projects. However, they hardly feature in the volume. It is also not entirely clear who the volume is aimed at. Nevertheless, the book offers a good introduction for those working in the cultural sector who are approaching the topic of AI for the first time. For advanced readers, however, this volume may offer too little.