Position talk 1: Nigel Guérin-Garnett
Title: Choreographing Through Experience: A Resistance
Bio:
After moving on from his dance career, Nigel worked in fashion design and development
roles across the Far East, Paris and New York. Following 15 informative years, he relocated to London for graduate studies at Kingston School of Art before embarking on a PhD at the Royal College of Art. Drawing on his hybrid background, these studies have helped to shape and reposition his practice into consulting and research. Straddling industry and academia, he has advised global luxury groups, multi-national corporations, creative festivals and independent SMEs through design research methods. His current areas of expertise include digital transformation, R&D and futures.
Abstract:
Contemporary dance offers a rich context to explore relationships through means beyond
linguistic or situational. Given its inherent corporeality and ephemerality, this artistic practice is well positioned to investigate the permeation of technological developments, such as artificial intelligence (AI), in society.
Focusing on the human body, my research considers the impact of intelligent digital agents
within creative processes, using choreography as a lens. Through the practice as research
methodology employed, I have developed a framework whereby non-human actors, such as
learning algorithm, may be implemented through more balanced means.
Developed iteratively, my novel choreographic system design assimilates artistically intuitive and digitally-based computational inputs to remotely communicate between
sender/researcher and interpreter/participant. It is hoped that the technologically affected material produced may promote questions about AI advancements within artistic pursuits, towards a rebalancing of the information economy. This research practice is positioned as a form of resistance, to challenge the increasing deployment of data collection technologies such as human gait recognition for corporate gain.
Position talk 2: Eva Cetinic
Title: Multimodal AI Models – Tools or Objects of Study?
Abstract:
The remarkable recent advancements of generative AI technologies, particularly their ability to perform convincing and inventive multimodal transformations (text-to-image, text-to-video, text-to-sound, etc.), marked a revolutionary milestone in AI development and triggered a strong interest not only within the AI community, but also among a broad scope of artists, researchers and practitioners fascinated with these technologies. The emergence of these technologies started to provoke discussions about the various implications of these technologies in the well as fundamentally challenged our perception of images and media content. As these technologies are now becoming widely integrated in many digital creative frameworks, it is becoming increasingly important to better understand the technological foundations and the various cultural, ethical and societal implications of the underlying AI foundation models. Trained on millions of image-text pairs sampled from the Internet. This is what makes them fascinating and problematic at the same time. These models are becoming widely used as tools for creation, but they can also be perceived as objects that need to be studied. Such models can be understood as repositories of cultural memories, which are capturing and reflecting collective experiences, dominant values and biases embedded in the vast amount of data they are trained on. Although we are becoming increasingly confronted with synthetic imagery produced by such models, the complex mechanism of associations encoded within those models remains unexplainable. While much attention has been devoted to generating realistic
become important to gain a better understanding of the underlying cultural foundations and the they have on the output of these models.
Bio:
Eva Cetinić is a researcher at the Digital Society Initiative, University of Zurich, where she is conducting her research project «From Text to Image with AI: How Multimodal Deep Learning Impacts Art and Culture». Before joining DSI, she was a postdoctoral fellow at the Center for Digital Visual Studies, University of Zurich; and a postdoctoral researcher in Digital Humanities and Machine Learning at the Department of Computer Science, Durham University, UK. Her research interest is focused on exploring deep learning techniques for computational image understanding and multimodal learning in the context of visual art and culture.
Position Talk 3: Mark d’Inverno
Abstract:
Stuart Rusell stated when thinking through the future of AI: “The right response seems to be to change the goals of the field itself; instead of pure intelligence, we need to build intelligence that is provably aligned with human values.” This has become known as the Value Alignment Problem and triggered a new body of research across ethics, social science and AI. I will explore how the value alignment problem can help us think through the role of AI in music. What are our values as people who make and listen to music? What are our values as performers and Improvisers? How does this help our approach to designing AI systems that help us to “sound right”?
Bio:
Mark d’Inverno is a critically acclaimed jazz pianist (BBC National Radio stations, jazz FM, BBC Music Magazine, Guardian, Observer, Jazz Review) who has performed nationally and internationally since the 1980s. He is an active research and teaching member of the academic community at Goldsmiths, University of London, having joined as a full professor in 2006. His interdisciplinary research considers the relationship between Artificial Intelligence and human creativity and spans the humanities, creative practice and pedagogical innovation. He has over 200 peer-reviewed publications, was formerly Director (Pro-Warden) of Research at Goldsmiths, and currently holds honorary professorial research positions at Instituto de Investigación en Inteligencia Artificial (IIIA) Universitat Autònoma de Barcelona, Spain and Dip. di Elettronica Informazione e Bioingegneria, Politecnico di Milano, Italy.