Bridging the Gap Between AI and Reality • Rhodes, Greece
Time: Wednesday, 5.11
Room: Room C
Authors: Cailean Finn
Abstract: The rise of Artificial Intelligence (AI) has produced agential systems that exhibit complex behaviours that cannot be fully explained or understood by breaking down the system into its subcomponents. As a result, for high-dimensional, opaque systems, the relationship between input and output can become so complex that their outputs are practically irreducible, which leads to a challenge in being able to predict the output of such systems. Instead, their behaviour is often a result of situated unpredictability, where outcome is stochastic but statistically bound. This unpredictability has the potential to lead to relational and emergent behaviours. Emergence can be understood as a complex phenomenon that is a product of coupled context-dependent interactions. It is the output of a global pattern of behaviour because of the local micro-interactions of actors, which cannot be explained through the sum of the individual agent’s behaviour. Traditions within Interaction Design and Human-Computer Interaction (HCI) have long argued for the value of designing for reflection, openness, and ambiguity as means of enriching experience beyond efficiency or control. Emergence extends these concerns by foregrounding indeterminacy not only as a feature of interpretation, but as a property of the system itself. In domains such as Generative Art and Artificial Life (A-Life), such behaviours are already treated as creative resources and prioritised goals, suggesting new possibilities for interaction design. The challenge lies in how emergence can be engaged as a design material within interactive systems. Emergence cannot be explicitly designed, since it exceeds specification. However, designers can design for emergence through boundary conditions, interaction rules, and feedback loops that create the possibility of unanticipated system-level behaviours. The role of design is not to script outcomes, but to shape the conditions under which novelty, richness, and relational complexity can arise and be meaningfully encountered. Prior work in HCI has emphasised how openness and ambiguity can invite reflection, prompt interpretation, and resist closure. Emergent systems can amplify these qualities, as it is defined by novel, unexpected and unpredictable outcomes, exceeding initial expectations, evoking feelings of surprise, and mystery through self-organisation and autonomy. While entanglement perspectives and related theories offer ways of conceptualising relational complexity, further work is required to explore how such perspectives can be translated into practices that directly address designing for emergence. In parallel, research in AI, Robotics, and A-Life have demonstrated emergence empirically in domains such as multi-agent systems, reinforcement learning, reactive robotics, and generative models. Such examples, illustrate how at a local and computational level, where complex interactions are defined by a series of often simple rules, these interactions can give rise to patterns of behaviour at the global or group level, because of the lower-level micro-interactions. This underscores the importance of emergence for understanding contemporary technologies enmeshed in a relational network of agency. However, its insights have not been systematically connected to design practices or evaluative frameworks within Interaction Design. While HCI has engaged related notions—such as openness, ambiguity, self-organisation, and relational complexity—the frameworks stop short of systematically theorising emergence itself. Work like Dourish’s embodied interaction, Kuuti & Bannon’s practice turn, or Hook & Lowgren’s strong concepts recognise indeterminacy and relationality, but they do not explicitly address how novel, system-level behaviours arise from local interactions across scales. By extending this trajectory, emergence is not only as a metaphor for openness, but material for design. This research addresses this gap through the creation of a series of interactive artworks that foreground emergence as a central design quality. Each work serves as a situated exploration of how emergent behaviours can be induced, encountered, and experienced in interaction. It also aims to demonstrates how practice-based inquiry can open new ways of designing for reflection, openness, and ambiguity by reframing indeterminacy as a constitutive and generative quality of interaction.