Performative AI: Human-Machine Agency
Chris Salter, Professor/Director Immersive Arts Space, Zurich University of the Arts (ZHdK).
Email: christopher.salter@zhdk.ch
Takashi Ikegami, Professor General Systems Science/Director Ikegami Lab, Graduate School, University of Toyko.
Email: ikeg@sacral.c.u-tokyo.ac.jp
This workshop will consist of an interdisciplinary roundtable of senior and emerging researchers from the sciences and arts discussing a new research project funded by the Swiss Federal Science Foundation involving researchers from Switzerland (Salter) and Japan (Ikegami). Our conception of Performative AI aims to reframe social-technical research and discussion around neural networks in order to explore such systems as “material agencies” that are temporally emergent and enacted by human and nonhuman actants (Latour 1996).
While “offloaded agency” between humans and machines (Salter and Ikegami 2023) has constituted a promising avenue for a renewed “systems aesthetics” (Burnham 1968; Penny 2017; Salter 2010), the mentioned theories grapple with a long-standing research issue across the human, natural and social sciences and the arts—that is, AI’s lack of “embodiment” (Merleau-Ponty 1962; Brooks 1991; Dreyfus 1992; Hayles 1999). Taking its cue from “enactive” or “embodied cognition” (Varela et al 1992; Pfeifer & Bongard 2006), long a mainstay in A-Life research, the workshop will engage with how embodiment as a minimal condition for intelligence subsides, despite or precisely because of today’s increased focus on “deep learning”-based neural networks (LeCun et al 2015), like Transformers (Vaswani et al 2017) that underpin Large Language Models (LLMs). However, the lack of physical grounding of LLMs poses significant challenges in understanding “qualia” (Nagel 1974). Therefore, relations between embodied action, language use and social-cultural-environmental settings in machine systems require further investigation (Koerner 1992) which the roundtable will discuss with examples from practice.
This workshop focuses on what it means for a system to display goal-directed behavior (GDB) , emphasizing how GDB can be described using tools from information theory, thermodynamics, and Bayesian statistics. Understanding GDB would help understand the gap between the seemingly purposeless universe of nonliving matter and the goal-filled world of life.
Panelists:
- Prof. Dr. Christopher Salter, artist and Professor/Director Immersive Arts Space, Zurich University of the Arts
- Prof. Dr. Takashi Ikegami, Prof. Physics, Department of General Systems Sciences, Graduate School of Arts and Sciences University of Tokyo.
- Dr. Anna Ciaunica, Centre for Philosophy of Science, University of Lisbon
- Yuko Hasegawa, Curator and Art Critic
- Dr. Alexander Saunier, Prof of Audio-Visual Arts (LUCA Belgium) and Senior Researcher, ZHdK
- Takahide Yoshida, PhD reseacher, Ikegami LabWhen is it justified to speak of a goal-directed system within a dynamical system for example a cellular automaton, a reaction diffusion system, or a chemical network?
7th International Workshop on Agent-Based Modelling of Human Behaviour (ABMHuB'25)
Dr Soo Ling Lim
University College London s.lim@cs.ucl.ac.uk
Prof. Peter J Bentley
University College London p.bentley@cs.ucl.ac.uk
This workshop aims to bring together researchers in agent-based modelling and agentic AI who are interested in using agent-based modelling to understand human behaviour. This combination of agent(ic)-based modelling and behavioural science is a growing area of research. Our motivation is to improve our understanding of collective human behaviour and address significant issues that are affecting the human population today, such as climate change, misinformation, opinion polarisation and global pandemics.
With large-language models (LLMs) providing entirely new ways to create agents and agentic architectures relating to human behaviour, we particularly welcome contributions in this area. Alife models offer the capability to create realistic laboratories for which to conduct experiments and progress our understanding in the area. We encourage researchers to use behavioural modelling to assess, challenge or even replace competing theories of human behaviour. Discussions of practical applications, ethical implications, and use cases from industry are also welcome.
Goal-Directed Behavior in Living and Non-Living Systems II
Richard Löffler,
University of Copenhagen,
Miguel Aguilera,
Basque Center for Applied Mathematics,
Martin Biehl,
Cross Labs, Cross Compass,
Omer Markovitch,
Instituto Superior Técnico,
This workshop focuses on what it means for a system to display goal-directed behavior (GDB) , emphasizing how GDB can be described using tools from information theory, thermodynamics, and Bayesian statistics. Understanding GDB would help understand the gap between the seemingly purposeless universe of nonliving matter and the goal-filled world of life.
We welcome submissions related GDB, including (but not limited to):
- What is the information content of goals and who or what makes use of this information?
- When is it justified to describe a nonliving physical system as exhibiting GDB?
- What are the minimal components driving GDB?
- Is it possible to construct systems, whether in vitro or in silico, that exhibit GDB? What is the role of GDB in life?
- How can goal-directedness emerge?
- When is it justified to speak of a goal-directed system within a dynamical system for example a cellular automaton, a reaction diffusion system, or a chemical network?
- Can goals be seen as secret keys to deciphering the behavior of systems?
Please check the workshop website for further information:
Workshop WEBSITEALife in Organizations: Navigating Societal Transitions Through Co-Creation
Gary Linnéuson gary.linneusson@his.se
Dept. of Industrial Eng. and Management, School of Engineering Science, University of Skövde, Skövde, Sweden,
Asimina Mertzani,
Dept. of Electrical & Electronic Eng., Imperial College London, London, UK
Nathan Lloyd,
Fac. of Business and Information Tech., Ontario Tech University, Oshawa, Canada
Modern organizations face increasing pressure to adapt to complex societal transition challenges — challenges that demand new forms of collaboration and decision-making. Climate change, demographic shifts, energy transitions, and digital transformation are driving the need for organizations to reconfigure their structures, goals, and strategies.
In this evolving landscape, organizations can no longer operate in isolation. Instead, they must improve their capacity to orient themselves in dynamic environments, balancing internal adaptability with increased awareness of their interactions with surrounding systems. Achieving this requires developing both:
- Micro-level (bottom-up) capabilities – empowering local teams, departments, and individuals to self-organize and respond to change in real time.
- Macro-level (top-down) capabilities – strengthening collective goals, systemic awareness, and strategic alignment to steer broader transformation.
Addressing these challenges requires new approaches that integrate insights from Artificial Life (ALife), complexity science, and cybernetics to navigate wicked problems, where we believe expanded understanding around concepts such as the following can be fruitful:
- Relational interfaces – the boundaries where internal organizational processes intersect with external societal systems. Expanding these interfaces is crucial to enhancing adaptability and unlocking new collaboration spaces across stakeholder groups.
- Co-creation arenas – emergent spaces that foster collaboration by identifying and connecting shared challenges, fostering collective goals, and creating conditions for constructive dialogue and innovation.
- Generative AI and Multi-Agent Systems – as potential tools to increase self-organizing capacities, enabling systems to dynamically adapt while balancing structure and flexibility.
We welcome submissions from diverse disciplines – including computer science, complexity science, organization theory, and systems thinking – as well as practitioners working on societal transition challenges. Our aim is to create a space for transdisciplinary dialogue, where participants collectively explore new strategies for enhancing adaptive capacity, maneuverability, and collaborative problem-solving in complex socio-technical environments.
Join us in Kyoto at ALife 2025 to explore how ALife principles can foster self-adaptiveness and co-creation in organizations – and beyond.
Emergence and Evolution of Language and Communication
Reiji Suzuki, Nagoya University (reiji@nagoya-u.jp)
Tadahiro Taniguchi, Kyoto Universit (taniguchi@i.kyoto-u.ac.jp)
Ryo Ueda, The University of Tokyo (ryoryoueda@is.s.u-tokyo.ac.jp)
The emergence of language and communication has been a major theme in artificial life research, beginning with studies on the evolution of communication protocols (Werner and Dyer 1991), naming games(Steels, 1995), and iterated learning (Kirby, 2001), with methodologies that have expanded across multiple fields including robotics, linguistics, cognitive science, evolutionary theory, philosophy, and complex systems sciences.
In recent years, the rapid advances of large language models have enabled artificial intelligence with sophisticated language processing capabilities, heightening expectations for AGI. However, numerous challenges remain concerning the origins of language – the source of"intelligence" – and its grounding in the real world.
To address these challenges, research on "Emergent Communication,"which approaches from a constructive perspective based on technological progress in deep learning and large language models, has gathered attention. Emergent Communication aims to model the process of language emergence and evolution. It is an interdisciplinary field where various research domains intersect, such as symbol emergence robotics, language evolution, and multi-agent reinforcement learning for communication.
This workshop aims to provide an interdisciplinary forum where researchers from various fields can gather to share and discuss the latest research findings on symbol emergence, emergence of communication, and language evolution. We hope to return to the foundational principles of artificial life research while integrating cutting-edge technologies. Through this, we seek to explore new research questions and strengthen cross-disciplinary collaboration.
Workshop WEBSITECultural Evolution of Planet X: Emergent Creativity and Wisdom of the Crowds
Dr. Suet Lee, University of Konstanz, Germany
Khulud Alharthi, University of Bristol, UK
Dr. Michael Chimento, University of Konstanz, Germany
Dr. David Garzón Ramos, University of Bristol, UK
Dr-Ing. Heiko Hamann, University of Konstanz, Germany
Prof. Sabine Hauert, University of Bristol, UK
Culture can be broadly defined as shared behaviours, values, beliefs, adaptive strategies, traits, or ideas that are passed from generation to generation via social learning. In particular, culture is not fixed but evolves over time in response to external and internal pressures, including transmission biases, cognitive transformation, and social structures and processes. Examples of cultures evolved in natural systems include vocal learning (e.g., bird or whale songs), nest building, tool manufacture and use, foraging behaviour, and movement.
Cultural evolution, therefore, is a powerful metaphor and framework to model and design adaptation mechanisms in complex systems across scales. It offers the potential to evolve the individual agents through quasi-global cultural knowledge. Falling under the branch of evolutionary algorithms, there are also intriguing links to artificial creativity and open-ended learning. For example, creative mediums such as music and language have evolved through the substrate of culture in animals.
We are interested in the relationship between biological and artificial cultures: can the cultural models evolved in robot swarms produce a plausible model for animal systems? On the other hand, can insights from how culture evolves in animal systems help us harness cultural evolution as a tool for learning and creativity in robot swarms? These are some of the questions we aim to highlight in an interactive and informal workshop setting. We hope to find synergies and spark collaboration between disciplines in the context of cultural evolution.
Workshop WEBSITEOrganizer: Tomas Veloz (tomas.veloz@vub.be)
Sponsored by:
- Centre Leo Apostel, VUB, Belgium;
- Institute of Advanced Studies/Chair of Transitions, UM6P, Morocco;
- Human Energy Project, USA;
- Departamento de Matematicas, UTEM, Chile.
We explore one of the core questions in ALIFE foundations: What is goal-directedness in complex systems. The particular perspective in this workshop is thinking this question through the interdisciplinary lens of matter-information interplays, and looking especially at metasystem transitions, which are the most radical forms of goal-directedness change. We aim to integrate what is typically discussed separately: e.g. biochemical models effectively explain cellular growth but overlook information processing, while information-centric approaches neglect physical embodiment.
Workshop WEBSITEGenerative Artificial Life: Simulation to Real World Computation
Takashi Ikegami, University of Tokyo
Hiroyuki Iizuka (iizuka@chain.hokudai.ac.jp), Hokkaido University
Artificial life research has approached the profound question of "what is life?" through philosophical, theoretical, and constructive methods. These efforts have revealed fundamental principles that under pin living phenomena, from self-organization and emergence to adaptation and evolution. Over decades of investigation, researchers have developed computational models, robotic systems, and chemical experiments that capture aspects of living systems, providing valuable insights into the mechanics of life processes.
However, despite these significant advances, we still cannot fully answer the straightforward question that lies at the heart of the field: how do we artificially create life? While we have made remarkable progress in simulating life-like behaviors and understanding individual mechanisms, the synthesis of these components into genuinely life-like systems remains elusive. The gap between our theoretical understanding and practical implementation continues to challenge researchers across disciplines.
This workshop deliberately addresses this simple yet unanswered theme of "how to create life," inviting presentations on approaches to life-creation from diverse perspectives. We welcome contributions from computer scientists modeling artificial ecosystems, chemists exploring protocell formation, engineers developing self-replicating robots, philosophers examining the conceptual foundations of life, and biologists investigating minimal living systems. By bringing together these varied approaches, we hope to identify common principles and complementary methodologies that could accelerate progress in the field.
Through these discussions, we will explore with participants what is missing from current approaches, what might be incorrect in our fundamental understanding, and how far we have progressed toward the goal of creating artificial life. We will critically examine whether our conceptual frameworks adequately capture the essence of life, whether our implementation strategies are sufficient, and what technological or theoretical breakthroughs might be necessary to bridge existing gaps.
By engaging in this dialogue across sub disciplines, we aim to discover potential breakthroughs that could unify the increasingly diversified artificial life community. As the field has expanded into specialized niches, opportunities for integration and synthesis have sometimes been overlooked. This workshop seeks to counter this fragmentation by fostering connections between researchers working at different levels of abstraction and in different material substrates, toward the tangible realization of life systems that embody the core properties we associate with living phenomena.
Workshop WEBSITE
.svg){kind=logo}
.png){kind=logo}
