Introduction: The Intersection of Metacreation and Cellular Automata

In contemporary creative practices, the paradigm of "Metacreation" has gained significant importance. This concept refers to a shift where the artist no longer determines the final form of a work directly but instead designs the "processes" or "systems" that generate the work. As Mitchell Whitelaw proposes, metacreation is a state where "the act of creation itself shifts toward the creation of process" . Within this context, Cellular Automata (CA) serve as the most iconic computational models embodying "Emergence"—where unpredictable complexity arises from simple local rules—playing a central role in both Artificial Life (ALife) research and media art .

This report redefines Cellular Automata through the lens of metacreation, investigating their mathematical foundations, evolution as artificial life, practical applications in media art, and their relationship with audience engagement (view counts and virality) on digital platforms. It specifically details the "discrete to continuous" evolution—from John Conway’s "Game of Life" to Bert Chan’s "Lenia"—and how recent integrations with foundation models and neural networks are rewriting the definition of creativity.

Mathematical and Theoretical Foundations of Cellular Automata

Cellular Automata are discrete computational models consisting of a regular grid of cells, each with a finite number of states, which update at discrete time steps according to fixed rules based on the states of neighboring cells.1 Originally conceived in the 1940s by Stanislaw Ulam and John von Neumann, these models were designed to describe the logic of self-reproduction.3

Components and Definition

A CA system is mathematically defined by the following elements:

  1. Lattice Dimension (): Usually 1D or 2D, though 3D models exist.4
  2. State Set (): A finite set of discrete states for each cell. In the simplest binary model, (representing "dead" and "alive").5
  3. Neighborhood: The spatial range of cells referenced to determine the next state. Common 2D neighborhoods include the "Von Neumann neighborhood" (4 adjacent cells) and the "Moore neighborhood" (8 surrounding cells).6
  4. Transition Rule (): A mapping that determines the next state based on the input from the neighborhood.6

Stephen Wolfram’s Classification of Complexity

In the 1980s, Stephen Wolfram classified the steady-state behavior of CA into four classes based on their complexity. This classification remains a key metric in metacreation for identifying rules that generate "life-like" diversity.8

Class Behavioral Characteristics Meaning in Metacreation
Class 1 Evolves to a homogeneous state (stagnation). Too static for creative processes.1
Class 2 Evolves into stable or periodic oscillating structures. Orderly, but limited in unexpected emergence.1
Class 3 Evolves in a seemingly chaotic, aperiodic fashion. High randomness; difficult to control patterns.2
Class 4 Emergence of complex structures that interact over long periods. Positioned at the "Edge of Chaos"; capable of universal computation.8

Wolfram suggested that the physical laws of the universe itself could be described using these simple rules, positioning metacreation not just as "imitation" but as the "construction of cosmic processes".9

Mitchell Whitelaw categorizes ALife-based artistic practices as "Metacreation," identifying four characteristic techniques. In this framework, Cellular Automata primarily belong to "Abstract Machines," used to investigate morphogenesis and pattern formation .

Four Categories of ALife Art

The technical tendencies of ALife art, organized from a metacreation perspective, are as follows:

  • Breeders: Use artificial evolution (genetic algorithms) where the artist selects and "breeds" individuals to generate forms. The artist's agency shifts from "painting" to "cultivating" .
  • Cybernatures: Construct complex, interactive artificial ecosystems that draw the audience into the system .
  • Hardware: Pursue embodied autonomy using "bottom-up" robotics .
  • Abstract Machines: De-emphasize biological analogy to investigate the mathematical logic of self-organization and pattern through models like CA .

These practices are based on the philosophy of viewing life not as a specific material form, but as a "flow" of information and "patterns of interaction".11 Christopher Langton noted that embedding "artificial molecules" into the logical universes of CA allows for the reproduction of the "molecular logic" of life .

From Discrete to Continuous: The Emergence of "Smooth" Life in Lenia

One of the most significant turning points in CA history was the 2015 introduction of "Lenia" by Bert Wang-Chak Chan. Lenia generalizes Conway’s "Game of Life" into continuous space, time, and states.6

Mathematical Generalization in Lenia

In Lenia, each cell state is a real value in the range $$ rather than a binary 0 or 1\. The time update follows a form similar to continuous differential equations 6:

where is the convolution with a kernel (neighborhood), and the growth function determines whether a cell increases (grows) or decreases (dies) based on neighborhood density.6 This transition allows Lenia to achieve extremely "organic and smooth" movements, resembling plankton or cells observed under a microscope.14

Digital Taxonomy and the Identification of "Species"

Bert Chan explored the vast parameter space of Lenia and identified over 400 "species" . These species exhibit symmetries, self-repair, locomotion, and even complex dynamics resembling feeding or fission.17

Family Key Characteristics Notes
Orbium Continuous "gliders" that move at a constant velocity. The most basic mobile entity in Lenia .
Scutium Shield-like forms that maintain a stable internal structure. An example of robust self-organization.14
Bizarre Cells Exhibit complex multicellular behaviors and irregular splitting. Gained significant attention on social media.13

Lenia won the "Virtual Creatures Contest" in 2018 and received the Outstanding Publication award from the International Society for Artificial Life (ISAL).6 This signifies that metacreation has moved beyond mere "simulation" to a phase akin to the "discovery of new life forms."

:::live lenia:::

Practical Applications of CA in Media Art

Cellular Automata have inspired many media artists due to their visual emergence and autonomy.

Ralf Baecker: Machinic Imagination

Ralf Baecker explores the aesthetics of computational generative processes. His work, Cellular Performances, is an audiovisual set using CA that pays homage to early computer programming while generating sound through real-time pixel scanning . Here, CA functions as an expression of "machinic imagination" where sound and image are inextricably linked.

Troika: Physical Algorithms

The London-based collective "Troika" translates digital algorithms into physical sculptures and drawings. Works like Life and Death of an Algorithm suggest that CA rules can extend beyond the canvas, using materials like aluminum and plastic to express the "vitality of computation".1 They position algorithms as powerful tools to express "fate and chance" rather than just mathematical exercises.19

Developments at ZKM (Center for Art and Media Karlsruhe)

ZKM is a global hub for algorithmic art. Artists like Bernd Lintermann and Ludger Brümmer have created numerous interactive installations using CA .

  • Morphogenesis (1997): A project by Lintermann simulating the development of organic objects .
  • CellularAutomataExplorer (2017): A tool by Brümmer that converts cell states into musical notes, generating highly complex sound structures .

These works highlight the core of metacreation: the artist designs the "rules for birth" rather than the specific melody or shape.

Takashi Ikegami Lab and the Frontier of Artificial Life

Professor Takashi Ikegami (University of Tokyo) leads ambitious projects extending CA logic to embodiment, consciousness, and social systems.

Integration of Alter3 and Large Language Models (LLMs)

The "Alter3" project features a humanoid robot equipped with GPT-4. It is not a simple automaton; it senses its own movements and feeds them back to the LLM, researching the moment "self" or "consciousness" emerges through the observation of its own signals.5

Significant milestones in Ikegami's activities (2024–2025) include:

  • Establishment of the ALife Lab in Kyoto (October 2025): A new base for deepening ALife research and social implementation.22
  • ALIFE 2025 International Conference (October 2025): Held in Kyoto with the theme "Ciphers of Life," integrating scientists and artists. Ikegami served as an advisor .
  • Venice Biennale 2025: Exhibition of an LLM-version of his 2010 work MTM (Mind Time Machine), presenting the "awakening" of consciousness at the intersection of massive data and embodiment.5

In Ikegami's philosophy, CA is redefined as part of a "massive data flow" where energy, matter, and information self-organize.9

Automating Search and Discovery: Sakana AI and ASAL

In late 2024, the Tokyo-based startup Sakana AI announced "ASAL (Automated Search for Artificial Life)," an algorithm that uses foundation models (FMs) to automatically "discover" ALife simulations . This marks a shift from "human design of rules" to "AI exploration of rules" (meta-metacreation).

Three Search Mechanisms of ASAL

ASAL uses the "human-aligned" internal representations of vision-language models as evaluation functions to perform three tasks 7:

  1. Supervised Target Search: Identifies parameters that match natural language prompts, such as "self-replicating patterns" or "things that look like real cells".7
  2. Open-Endedness Search: Searches for simulations that generate "temporally open-ended novelty," discovering worlds that are as "expressive" as Conway’s Game of Life.7
  3. Illumination: Maps the parameter space to identify a diverse set of interestingly different phenomena (e.g., various Lenia species).2

ASAL is substrate-agnostic and works across Boids, Particle Life, and Neural Cellular Automata, completing "interesting parameter searches" in seconds that previously took humans months.23

Digital Popularity: View Counts and Visual Psychology

The phenomenon of CA videos, particularly Game of Life and Lenia, recording explosive view counts (virality) on YouTube and SNS suggests the psychological impact of metacreation on the public.

Why CA Goes "Viral": A Visual Psychology Approach

CA content is compelling due to the cognitive structures of the human brain:

  • Edge of Chaos: Humans are drawn to patterns that are neither purely ordered (monotonous) nor purely chaotic (random), but at the boundary where "unpredictable meaning" exists. Wolfram's Class 4 behavior hits this psychological "sweet spot" .
  • Biological Affinity: The brain's visual cortex (e.g., area V1) contains specialized cells for detecting edges and textures. The formation process of CA patterns resonates with the brain's own neural dynamics.30
  • "Satisfying" Content: Many viewers find Lenia’s smooth movements or the self-repair of CA "satisfying" or "healing." This is because the fundamental brain functions of pattern recognition and information compression trigger reward systems.3

Success in CA content is backed by specific data trends:

  • Extreme Game of Life: Constructions of CPUs or complex machines evoke intellectual curiosity, described by viewers as "narratively rich" or like "reading fantasy lore" .
  • Lenia/Organic CA: Smooth moving entities and fission processes trigger emotional responses, with viewers noting it looks like "real living creatures" .

Strategic use of tags (Generative Art, ALife, Satisfying) and comparing performance against successful competitors are key strategies for increasing engagement .

As metacreation entrusts creation to "autonomous systems," the question of "authorship" and rights attribution becomes unavoidable.

Current copyright laws are "anthropocentric," a premise challenged by metacreation.

  • Rejection of AI Copyright: In 2025, U.S. federal courts (e.g., Thaler v. Perlmutter) reaffirmed that works created autonomously by AI cannot be copyrighted, as human authorship is a foundational requirement .
  • Defining the Author: In AI art, the "author" can be defined in four ways, each leading to different legal conclusions: 1\) the prompter, 2\) the programmer, 3\) the AI itself, or 4\) a collaborative assemblage .

Audience Perception of Value

Research indicates that audiences tend to rate works higher when they believe they were "created by a human" rather than an AI (Mean rating: Human 5.72 vs. AI 4.99) . This reflects the persistence of the traditional value that art is an expression of human emotion. However, proponents of metacreation argue that the value lies in "conceptualizing the system" and "providing the field for emergence" rather than the final output .

Conclusion: Future Outlook for Metacreation

The investigation of metacreation through Cellular Automata leads to fundamental questions about how we define life, intelligence, and creativity. From simple grid flickers to the smooth life forms of Lenia, and now to automated discovery via Sakana AI’s ASAL and Ikegami’s embodied AI (Alter3), CA has evolved significantly.

  1. Integration of Discrete and Continuous: The success of Lenia re-integrated "digital (discrete)" descriptions of the world with "analog (continuous)" appearances. Future systems will likely build even higher-dimensional "smooth worlds".13
  2. Autonomous Discovery by AI: We are moving from an era where humans create rules to one where AI explores "interesting rules" for humans to interpret. Systems like ASAL will drastically accelerate scientific discovery.7
  3. ALife as Social Infrastructure: As Ikegami suggests, ALife theories will be applied to web data flows, urban traffic, and the design of consciousness. Metacreation provides a new "Generative Ethics" for designing social systems.23

Cellular Automata serve as a "mirror" in metacreation, allowing us to describe and experiment with what we want life to be. The increase in view counts and viral spread on social media is a modern expression of our fundamental wonder at the "vitality" reflected in that mirror.

---

Comparison of Key CA and Metacreation Software/Libraries

Software Developer/Org Key Features Target
Golly 5.0 Andrew Trevorrow, etc. World-renowned CA explorer. Uses Hashlife for extreme speed. Discrete CA, Game of Life 4
Ready Golly Gang Handles continuous reaction-diffusion and FDTD models. Continuous CA
CAX Open Source (JAX) GPU/TPU-accelerated CA library. Differentiable. NCA, Lenia, Large experiments 33
ASAL Sakana AI Foundation model-driven automated discovery system. All ALife substrates 2
Lenia Explorer Bert Chan Interactive browser-based explorer for Lenia species. Lenia, General users 13

引用文献

  1. Trickle Down Aesthetics – Cellular Automata in recent works by Troika, 5月 10, 2026にアクセス、 https://www.creativeapplications.net/project/trickle-down-aesthetics-cellular-automata-in-recent-works-by-troika/
  2. Automating the Search for Artificial Life with Foundation Models \- Sakana AI, 5月 10, 2026にアクセス、 https://sakana.ai/asal/
  3. Cellular Performances, cellular automata audiovisual set \- Neural, 5月 10, 2026にアクセス、 https://neural.it/2025/11/cellular-performances-cellular-automata-audiovisual-set/
  4. Golly Game of Life Home Page, 5月 10, 2026にアクセス、 https://golly.sourceforge.io/
  5. What are good examples of cellular automata in art? \- Quora, 5月 10, 2026にアクセス、 https://www.quora.com/What-are-good-examples-of-cellular-automata-in-art
  6. Lenia \- Wikipedia, 5月 10, 2026にアクセス、 Lenia" target="_blank" rel="noopener noreferrer">https://en.wikipedia.org/wiki/Lenia
  7. Automating the Search for Artificial Life with Foundation Models, 5月 10, 2026にアクセス、 https://asal.sakana.ai/
  8. Cellular Automata \- Artificial Life, 5月 10, 2026にアクセス、 https://alife.org/encyclopedia/introduction/cellular-automata/
  9. Cellular Automata in Architecture, Design, and Art | by Stepan Kukharskiy | Medium, 5月 10, 2026にアクセス、 https://medium.com/@stepan.kukharskiy/cellular-automata-in-architecture-design-and-art-217fce9c1a5c
  10. Can cellular automata be a representative model for visual perception dynamics? \- PMC, 5月 10, 2026にアクセス、 https://pmc.ncbi.nlm.nih.gov/articles/PMC3787243/
  11. STUDYING ARTIFICIAL LIFE WITH CELLULAR AUTOMATA\* Christopher G. LANGTON \- Gwern.net, 5月 10, 2026にアクセス、 https://gwern.net/doc/cs/cellular-automaton/1986-langton.pdf
  12. ALIFE 2025, 6–10 Oct, Kyoto, Japan | Ciphers of Life \- International Society for Artificial Life, 5月 10, 2026にアクセス、 https://2025.alife.org/
  13. Lenia \- Bert Chan, 5月 10, 2026にアクセス、 [lenia.html](https://chakazul.github.io/lenia.html)" target="_blank" rel="noopener noreferrer">https://chakazul.github.io/lenia.html](https://chakazul.github.io/lenia.html)
  14. Lenia: Biology of Artificial Life by Bert Wang-Chak Chan \- Complex Systems, 5月 10, 2026にアクセス、 https://www.complex-systems.com/abstracts/v28\_i03\_a01/
  15. Stanford Seminar \- Lenia: Biology of Artificial Life, Bert Wang-Chak Chan \- YouTube, 5月 10, 2026にアクセス、 https://www.youtube.com/watch?v=G5P8eu6gUSo
  16. The Ownership of AI Art: Cultural Sustainability, Ethical Authorship, and Museum Governance in AI-Assisted Artistic Practices \- Preprints.org, 5月 10, 2026にアクセス、 https://www.preprints.org/manuscript/202601.2376
  17. Lenia \- Artificial Life, 5月 10, 2026にアクセス、 [lenia/](https://alife.org/encyclopedia/software-platforms/lenia/)" target="_blank" rel="noopener noreferrer">https://alife.org/encyclopedia/software-platforms/lenia/](https://alife.org/encyclopedia/software-platforms/lenia/)
  18. Nobody Can Agree on What AI Art Actually Is — and That's a Problem | by Kejama \- Medium, 5月 10, 2026にアクセス、 https://medium.com/@luckykejama/nobody-can-agree-on-what-ai-art-actually-is-and-thats-a-problem-904881b40a56
  19. Creativity reimagined: ethical authorship, ownership and social impact of AI-generated art, 5月 10, 2026にアクセス、 https://www.emerald.com/jices/article/24/2/250/1303557/Creativity-reimagined-ethical-authorship-ownership
  20. Takashi Ikegami Lab / 池上高志研究室 \- 東京大学, 5月 10, 2026にアクセス、 https://www.sacral.c.u-tokyo.ac.jp/
  21. Future Ideations Camp Vol.5 : Can AI Become Life? \- Civic Creative ..., 5月 10, 2026にアクセス、 https://ccbt.rekibun.or.jp/en/events/future-ideations-camp05
  22. Cellular automaton \- Wikipedia, 5月 10, 2026にアクセス、 https://en.wikipedia.org/wiki/Cellular\_automaton
  23. 人工生命研究者・池上高志が考えてきたこと \#4 \- DISTANCE.media, 5月 10, 2026にアクセス、 https://distance.media/article/20250922000533/
  24. PUBLICATIONS \- Takashi Ikegami Lab / 池上高志研究室, 5月 10, 2026にアクセス、 https://www.sacral.c.u-tokyo.ac.jp/publications
  25. Automating the Search for Artificial Life with Foundation Models \- arXiv, 5月 10, 2026にアクセス、 https://arxiv.org/html/2412.17799v1
  26. Neural Cellular Automata For Art Recreation \- Towards Data Science, 5月 10, 2026にアクセス、 https://towardsdatascience.com/neural-cellular-automata-for-art-recreation-6d9fb61afb37/
  27. Cellular Automata: Self-reproduction of Order and Beauty | by Dorit Geifman | Medium, 5月 10, 2026にアクセス、 https://medium.com/@dorit.geifman\_38625/cellular-automata-self-reproduction-of-order-and-beauty-e60082ddec6f
  28. SakanaAI/asal: Automating the Search for Artificial Life with Foundation Models\! \- GitHub, 5月 10, 2026にアクセス、 https://github.com/sakanaai/asal
  29. Why the Obsession with Human Creativity? A Comparative Analysis on Copyright Registration of AI-Generated Works \- HLS Orgs, 5月 10, 2026にアクセス、 https://orgs.law.harvard.edu/halo/2025/02/21/why-the-obsession-with-human-creativity-a-comparative-analysis-on-copyright-registration-of-ai-generated-works/
  30. YouTube Video Analytics Checker \- Analyse Any Video \- Lenos, 5月 10, 2026にアクセス、 https://www.lenostube.com/en/youtube-video-analytics-checker/
  31. Algorithmic Art \- Google Arts & Culture, 5月 10, 2026にアクセス、 https://artsandculture.google.com/story/algorithmic-art-zkmkarlsruhe/2gXBp4XkgyBNCQ?hl=en
  32. Ethical dilemmas in artificial intelligence-generated art: authorship, ownership, and the blurring of creative boundaries | Digital Scholarship in the Humanities | Oxford Academic, 5月 10, 2026にアクセス、 https://academic.oup.com/dsh/advance-article/doi/10.1093/llc/fqag035/8554352?searchresult=1
  33. simulation \- Software for cellular automota \- Computational Science Stack Exchange, 5月 10, 2026にアクセス、 https://scicomp.stackexchange.com/questions/30835/software-for-cellular-automota