The Showcase method creates the setting of a simulated curation or fictional showcase to encourage exploration and communication of the state of the art of machine learning and its future implications for the public. The method consists of four parts following different approaches and mind sets, from design thinking in particular and creative processes in general: exploring (discover / look), creating (define / think), prototyping (develop / make) and presentation (deliver / share).
Name: The Showcase
Partner organization: Technische Universität Dresden
Authors: Robert Fischer, Henriette Greulich, Daniel Lordick, Lisa Nickolaus
Aims: This method enables collaborative work on a specific topic with people from different professional and disciplinary backgrounds and aims to bridge the gap between scientific, artistic and societal perspectives and concepts related to the workshop’s topic.
Time: We recommend a duration of at least five hours.
Original Context: This method was developed for a two-day online workshop Artificial Intelligence Exchange (AIX), at the Technische Universität Dresden. The participants worked on a showcase for artificial intelligence in groups of up to eight persons. The results of the group process were to support the communication of AI concepts to the public. Besides developing a concept for the showcase, lectures on art and AI, on machine learning and on linguistics and AI were included, framing the Showcase method.
Group: The method was tested with a group of students, PhD candidates, postdocs and professors from different backgrounds (mathematics, social sciences, computer science, art, physics). The lectures were given by two scientists (a Professor of Linguistics and a Research Associate for Software Development) and two artists in residence at TU Dresden (who are currently working at the intersection of art, artificial intelligence and society).
A collaborative digital platform that enables creative multimedia group work and communication via video. During our workshop we used a combination of MIRO and Zoom. We also used Mentimeter as an online tool for creating and using real-time feedback.
To playfully explore the field of artificial intelligence, and in particular machine learning, we also offered a collection of contemporary tools, e.g. for:
- Text analysis and text generation – GPT-3 (trial accounts for every participant with up to 300 K tokens each)
- Image processing – GAN Playground (Generative Adversarial Networks);
- Music generation – MuseNet.
(Visit website for more details)
To work with 30 to 40 participants we recommend dividing into four or five groups and guiding with one moderator per group who should be familiar with Design Thinking and STEAM methods. Additionally, one moderator leads the workshop, clarifies the tasks of the Showcase method and moderates the final presentations and panel discussion. There should also be someone to prepare and guide the Zoom sessions. Lecturers or experts with diverse backgrounds are required to present multiple perspectives on the topic, provide guidance through the AI software applications, and to answer questions that arise during the workshop.
Potential application/ adaptation: The method is transferable to topics other than artificial intelligence.
This method enables collaborative work on a specific topic with people from different professional and disciplinary backgrounds and aims to bridge the gap between scientific, artistic and societal perspectives and concepts related to a specific topic. The goal of the original application of the method was to explore and be able to communicate in a particular way the state of the art of machine learning and its future implications to the public along thematic lines, such as problem solving, business concepts, creativity, memory and individuality. Facing this, art, design and curation can open up a range of strategies to connect and communicate different perspectives which the Showcase method utilizes. Its task is to create, prototype and present a concept of a (fictional) showcase related to a specific topic (e.g. AI) with the opportunity to be presented on a (real) website.
To this end the following questions were asked:
- What aspects of AI am I fascinated/threatened by?
- What aspects are the most interesting/challenging for me/our modern society?
- How can I translate at least one aspect of AI into a physical/digital/conceptual output for public exhibition?
- How can I communicate aspects of AI for a broader public (as a scientist, artist, research assistant, company, student, teacher, curator, etc.)?
The Showcase method consists of four parts following different approaches and mind sets from design thinking in particular to creative processes in general:
- first phase of exploring (discover / look)
- second phase of creating (define / think)
- third phase of prototyping (develop / make)
- fourth phase of presentation (deliver / share).
We recommend guiding the group work with one moderator per group who should be familiar with STEAM and Design Thinking methods and be able to offer selected methods from a pool of methods for the group, if methodological support is needed during the process.
Experimenting I and II- Time to explore and create!
Due to organizational reasons we subdivided this frame into two parts stretching over the two days. The basic idea was to be hands on with the AI software tools, play with them and get to know their strengths and weaknesses, possibilities and limits. In parallel with the experimentation, the groups started to shape a concept for the AI showcase, with methods provided as needed.
Prototyping – Time to make!
Prototyping methods like Videocut, Wire Framing and (Meta-) Storyboard were used to further shape the concept of the group’s showcase and prepare a presentable output (not necessarily physical) for the following pitch.
Final Pitch – Time to present!
The groups were given three minutes each to briefly present their understanding of AI and prepared concept of the AI showcase (pitch).
Afterwards, all of the participants were invited to ask questions and to feed a prepared Mentimeter with the most outstanding impression / concept / idea of the pitch. Mentimeter is an online tool for creating presentations and including real-time feedback.
Introductory tasks like the Future Timeline and lectures for creating a common knowledge base, as well as inspiring the participants, are recommended. Furthermore, a short briefing on STEAM methodology (and additionally Design Thinking concepts) is mandatory. The original application of the Showcase method was built on two talks by artists (who are currently working on topics at the intersection of art, artificial intelligence and society) and on two talks by scientists (a Professor of Applied Linguistics and a Researcher at the Chair of Software Technology for software development). The scientists and experts also provided an introduction to (digital) tools of interest relating to different applications of machine learning.
As a follow-up a panel discussion regarding outcomes and new findings as well as a reflection on the methods used is recommended. A closing panel discussion can include both the session’s lead moderator and the lecturers (artists and scientists), who can share their view on the outcomes, and also participants, who could assess their new findings, constructive annoyances and perceived knowledge gained. A final reflection on the methods and the possible posting of the showcases on a website are other options for disseminating the outcomes provided participants agree to this. Part of the reflection and discussion phase could be the visualization of real-time feedback with a Mentimeter as described above.
(source: Applying Design Thinking: A Workbook for Acadamics and Researchers in Higher Education, URL: https://tu-dresden.de/ing/maschinenwesen/imm/td/ressourcen/dateien/forschung/eBook_2_0_English.pdf , p. 16)
The Showcase MIRO board overview: Each column belongs to one of the four phases and each row belongs to one of the four groups.
One group’s process and product is shown as an example in the following frames structured according to the four phases of the Showcase method.
Text included in the presentation phase:
“When we talk about machines and artificial intelligence, we often only think about consciousness.
However, according to psychoanalysis, consciousness is just the tip of the iceberg; there is a whole submerged world, mostly unknown, which we still have to explore. Some unconscious contents drive us throughout our daily lives, they feed our creativity, and can re-emerge in our dreams; they come to us during the sleep, to be seen and understood.
Sometimes what the dream experience leaves within us can be an image, a sequence of images, a story, sounds, or sensations.
With AI we can now write down what we remember, generate images and sounds from it, and even continue the dream; with AI we can re-create the video-story of our dreams.
By better understanding our inner universe, we can improve both our lives and societies.”
Another group’s process and product is shown here in the following frames structured according to the four phases of the Showcase method.
Templates for integrating additional methods as well as the methodological concept were provided on a second MIRO board (only accessible to the moderators) as shown below. They could easily be used with drag and drop or copy and paste.
The workshop was considered a success primarily if the groups were able to share ideas and develop them collaboratively into one AI showcase. Therefore, the focus was set primarily on the process and not on the result itself. Nevertheless, we encouraged the participants to develop a result by offering the possibility to publish the showcase on a website.
We evaluated the workshop and its components with the tool Mentimeter and through an in-group discussion.
Feedback was overall positive with some suggestions for improvements.
Participation was voluntary and the showcases were not tested.