The “Una Europa Summer School” brought together doctoral researchers from across Europe to explore the intersections of artificial intelligence, creativity, and societal innovation. Organized within the framework of Una Europa, an alliance of leading European universities committed to fostering interdisciplinary collaboration and European cooperation in higher education and research, the program offered participants opportunities to attend workshops, engage in discussions, and collaborate on projects addressing contemporary challenges and future developments in AI.

As part of the Una Europa Summer School’s topic on “AI & Creativity” at University College Dublin, around 25 PhD candidates from diverse disciplinary backgrounds participated in the Magic AI Machine Workshop, inspired by Andersen and Wakkary (2019), an interactive and creative session led by Peter Sörries and Jacobe Klein from the Human-Centered Computing Research Group at the Institute of Computer Science, Freie Universität Berlin. The workshop invited participants to rethink artificial intelligence through the lenses of imagination and speculation. Rather than focusing on existing AI technologies or technical implementation, the session encouraged participants to critically explore alternative and preferable futures for AI. Drawing inspiration from speculative design, participants were challenged to create visions of AI that do not yet exist, but perhaps could.

Activity

The workshop began with an introduction to speculative design, an approach that uses imaginative, often fictional scenarios to provoke discussion of future possibilities and societal implications. Participants were introduced to the concept of the Future Cone of Possibilities, which illustrates how different futures, namely possible, plausible, probable, and preferable, can emerge depending on technological developments and social, political, and ethical choices. Building on these concepts, participants formed six interdisciplinary groups and engaged in a hands-on design activity. Using a wide range of craft materials, they were tasked with creating their own “Magic AI Machines,” which are physical artifacts representing future AI systems designed to address societal challenges and strengthen human resilience. The central design prompt asked: How can we overcome the black-box problem of AI—what AI is, how it works, and how decisions are made—to become more resilient as individuals and societies? Importantly, the activity prioritized imagination and critical reflection over technical feasibility. Participants were encouraged to move beyond current technological constraints and envision AI systems that promote transparency, human values, and collective well-being.

Selected Outcomes

The workshop culminated in a gallery walk and presentation session, where participants used storytelling and role-play to showcase their speculative machines. The resulting machines demonstrated a remarkable diversity of perspectives on how AI could become more understandable, accountable, and socially beneficial.

For example, “The Looking Box,” a machine designed to reveal AI processes through different human perspectives and needs. By applying various metaphorical filters, users could “look inside” otherwise opaque AI systems, highlighting how understanding AI depends on individual experiences, contexts, and values.

Another group developed “Transcending the Black Box,” featuring a cloud-shaped structure representing AI knowledge. Rather than depicting intelligence as residing solely within machines, the concept emphasized that AI knowledge continuously grows through networks of human-AI interactions, collaboration, and shared learning. The machine challenged conventional notions of AI as an isolated technology, instead framing it as a socio-technical system co-created by people.

The “X-Plain” machine focused on making AI data flows more visible and easier to understand. By representing different data streams and processing stages, the machine sought to demystify how information is collected, transformed, and used by AI systems. By making these processes tangible, the concept aimed to improve data literacy and strengthen users' capacity to critically engage with AI-driven decisions.

“On the Way to Balance” took the form of a robot traveling along a long and winding road. The machine symbolized the ongoing journey towards balancing data-driven innovation with human needs, ethical considerations, and societal values. Rather than presenting balance as an achieved state, the design highlighted it as a continuous process that requires negotiation, reflection, and collective responsibility.

Another group presented “ORECI,” a framework for transparent and ethical AI interaction. The concept envisioned AI systems that actively communicate their intentions, limitations, and decision-making processes. By placing human values at the center of AI interactions, the machine encouraged participants to consider how trust and accountability might be designed into future technologies from the outset.

Reflections

Beyond the individual machines, the workshop generated lively discussions about the opportunities and challenges of creating more transparent, explainable, and value-sensitive AI systems. Many of the concepts acknowledged that significant technological, ethical, and societal hurdles remain. Participants, for example, reflected on what future AI systems might look like and on the pathways and collective efforts required to realize them. A central takeaway from the workshop session was that designing AI technologies is also about designing human experiences, relationships, and futures. By materializing speculative futures through tangible artifacts, participants were able to critically examine AI's role in society and articulate the kinds of futures they would prefer to create.

References
Kristina Andersen and Ron Wakkary. 2019. The Magic Machine Workshops: Making Personal Design Knowledge. In Proceedings of the 2019 CHI Conference on Human Factors in Computing Systems (CHI '19). Association for Computing Machinery, New York, NY, USA, Paper 112, 1–13. https://doi.org/10.1145/3290605.3300342