A panel discussion about laboratories for artificial intelligence and machine learning in design education and design research with special guests Alexa Steinbrück and Simon Maris, who run the XLab at Burg Halle, as well as all KITeGG Lab Leads.

Reflections on 10+ years of teaching machine learning to creative practitioners

For designers and other creative practitioners, learning to effectively reason about and apply machine learning can open up new creative possibilities, fruitful interdisciplinary collaborations, and new ways of working. However, traditional machine learning curricula and learning resources usually aren’t a good fit for the types of knowledge and skills that creative practitioners need. In this talk, I’ll discuss my experiences in over more than a decade of teaching machine learning to creators, highlighting some of the approaches I’ve found to be effective as well as some of the persistent challenges, and encouraging others to formalise and share their own approaches to teaching.

Rebecca Fiebrink s a professor at the Creative Computing Institute at University of the Arts, London and affiliated with the Department of Computing at Goldsmiths University. Her teaching and research focus largely on exploring how machine learning can support and expand human creative practices. She is the developer of the »Wekinator« creative machine learning software, and she created the first online class about machine learning for music and art.

The End of AI Art

Recent multi-modal generative models like DALL-E 2 or Imagegen have attracted much attention, and have been perceived as both a significant next step in, and radical disruption of, the automated creation of non-trivial images. In this talk, I will trace the development of these techniques over the past decade and argue that we have reached the end of AI art. The photorealism of current-generation models is a stock-photorealism. As model-specific glitches and constraints (e.g. mode collapse in GANs, compression artifacts in VAEs) are eliminated, and as massive meta-datasets of Internet-vernacular imagery become the new norm, the aesthetic contact surface of machine learning vanishes. While this does not limit the theoretical potential of machine learning, it does mark the end of a distinguished period of experimental practice.

Fabian Offert is Assistant Professor in History and Theory of Digital Humanities at the University of California, Santa Barbara. Before joining the faculty at UCSB, Fabian was Postdoctoral Researcher in the DFG priority program “The Digital Image” at Friedrich Alexander University Erlangen-Nuremberg and Affiliated Researcher in the Artificial Intelligence and Media Philosophy Research Group at Karlsruhe University of Arts and Design.

☑ I’m not a …

Robots are no longer just “better” or more “efficient” workers, they also become co-actors in our daily lives. In our coexistence we ascribe almost human qualities, emotions and liveliness to our technical counterpart, but the rational machine is something else. But how might these things be designed if they evolve from trivial tools to proactive things? The talk focuses on the non-anthropomorphic appearance and behaviour of robotic others as an alternative approach to support interaction with non-human beings.

Andreas Muxel is Professor for Physical Human-Machine Interfaces at the Faculty of Design at University of Applied Sciences Augsburg where he founded and is directing the Hybrid Things Lab, exploring alternative perspectives to our relationship with proactive, robotic beings in terms of acceptance and trust. His contributions have been internationally published, exhibited and awarded (p.ex. ACM DIS Eindhoven, Ars Electronica Festival Linz, FILE Festival São Paulo, TodaysArt Festival Brussels, VIDA Award Madrid, Share Prize Torino). His research is centered around poetical and engaging tools for thinking, regardless of whether they are hardware or software.