Reimagining Nature Through Creative AI
fuse* (info@fuseworks.it)1
Metadata
Urban Eidos Volume 5 (2025), pages 26-39
| Journal-ISSN: | 2942-5131 |
| DOI (PDF): | https://doi.org/10.62582/UE5003p |
| DOI (online): | https://doi.org/10.62582/UE5003o |
Abstract – Over the past seven years, fuse* has undertaken a continuous exploration of botany, ecology, and artificial intelligence. Through three distinct but interconnected series – Artificial Botany, Unseen Flora, and Mimicry – the studio’s practice investigates how algorithmic tools can be used not only to simulate but also to reimagine natural systems, species, and evolutionary trajectories. From early experiments with botanical illustrations reinterpreted through neural networks to the speculative creation of hybrid plant-insect organisms, each project expands the narrative potential of artificial intelligence in relation to environmental themes, exploring each of them through creative lenses that combine scientific research and visual experimentation.
Artificial Botany
Artificial Botany (2019 – ongoing) is an ongoing project which explores the latent expressive capacity of botanical illustrations through the use of machine learning algorithms.
Before the invention of photography, botanical illustration was the only way to visually record the many species of plants. These images were used by physicists, pharmacists, and botanical scientists for identification, analysis, and classification. While these works are no longer scientifically relevant today, they have become an inspiration for artists who pay homage to life and nature using contemporary tools and methodologies. Artificial Botany draws from public domain archive images of illustrations by the greatest artists of the genre, including Maria Sibylla Merian, Pierre-Joseph Redouté, Anne Pratt, Marianne North, and Ernst Haeckel.
Developing as an organism in an interweaving of forms that are transmitted and flow into each other, the plant is the symbol of nature’s creative power. In this continuous activity of organising and shaping forms, two opposing forces in tension are confronted: on one hand, the tendency to the shapeless, the fluidity of passing and changing; on the other, the tenacious power to persist, the principle of crystallisation of the flow, without which it would be lost indefinitely. In the dynamic of expunction and contraction that marks the development of the plant, beauty manifests itself in that moment of balance, which is impossible to fix, caught in its formation and already in the point of fading into the next one.
Artificial Botany responds to the need to describe the creative power of nature both in visual and conceptual terms, restoring the concept of mutability, transience and evolution as basic elements of life. This multidisciplinary series combines digital and analogue prints with immersive video installations: a series so multifaceted that brings out the metamorphic characteristic of existence – everything changes constantly.
Creation Process
The creative process at the foundation of Artificial Botany is based on a particular machine learning system called GAN (Generative Adversarial Network). Through a training phase using botanical illustrations, the system can recreate new artificial images with morphological elements extremely similar to the initial figures: not an exact copy of the original image, but a reinterpretation of it. The machine re-elaborates the content by creating a new language, capturing the information and artistic qualities of man and nature.
GANs are made up of two networks that compete with one another in a zero-sum game framework: the first network is called a generator and its job is to generate data from a random distribution. These data are then conducted to the second network, the discriminator: based on the data acquired during the learning phase, it learns to decide whether the distribution of the generator data is close enough to what the discriminator knows as the original data. If the value generated does not meet the requirements, the process will be repeated until the result is obtained. GANs typically run unsupervised, teaching themselves how to mimic any given distribution of data – meaning that once trained they can replicate novel content starting from a specific dataset.
Figure 2: A visual outline of GANs.
The first step in establishing a GAN is to identify the desired output and gather an initial training dataset based on those parameters. This data is then randomised and input into the generator until it acquires basic accuracy in producing outputs.
In an unconditioned generative model, there is no direct control over the model and the data being generated. However, by conditioning the model on additional information it is possible to direct the data generation process. Such conditioning could be based on class labels, on some parts of data for painting-like features, or even on data from different modalities.
This system can be put in place infinitely, each time obtaining a distinct result that reflects the starting dataset. Artificial Botany is thus an extremely versatile project, whose aesthetic and visual identity get redefined and transformed according to the dataset at the basis of the GAN elaboration.
Figure 3: first tests from Artificial Botany.
Datasets and Archives
One of the main datasets used in the project consists of high-resolution images from both the illustrated herbarium and the dry herbarium of Ulisse Aldrovandi (1522 – 1605), the renowned Bolognese botanist and naturalist widely regarded as the father of modern Natural History. This was made possible thanks to the concession by BUB (University Library of Bologna), the Botanical Garden of Bologna and Alma Mater Studiorum.
By integrating the images of Aldrovandi’s illustrations in the processing of Artificial Botany, a peculiar and unique version of the Aldrovandian botanical and imaginative sample collection was generated: a modern exploration of the original illustrations that allows to create relations between stylistic elements and details that would probably pass unnoticed by a human eye.
Another iteration of the project is the synthetic reinterpretation of the collection of plants and floral illustrations of the Botanical Garden of Padua, the oldest university botanical garden in the world. The resulting work is Artificial Botany .erbario assoluto, a three-channel video installation acquired by the Garden and permanently exhibited in its newly-opened spaces: a unique crystallisation of the museum collection that puts in contact historical scientific discovery with technological innovation.
In this case, the project involved working with the digitized archive of a vast dataset of botanical illustrations from the library of Giovanni Marsili (1727-1795), prefect of the Botanical Garden of Padua who significantly contributed to the garden by giving life to an extensive book collection, which grew to include over 2,500 pieces from across Europe. The collection also includes peculiar pieces, first and foremost those created with the ‘smoke printing’ technique – exposing a dried plant to smoke and subsequently impressing its shape on a sheet.
Artificial Botany dialogues with the beauty implicit in the state of continuous transformation of living species, seeking to capture the generative richness of evolution just like it has been done in the past centuries through herbariums by botanists and scientists. Artificial Botany .erbario assoluto wants to imagine new relationships and assonances between the different species preserved in the Garden and their representations.
Figure 5: frames from Artificial Botany .erbario assoluto
Figure 6: installation view from Artificial Botany .erbario assoluto at Smart Life Festival, Modena (IT), 2023
Unseen Flora
One of the most recent iterations of Artificial Botany marked a significant departure from the original series, evolving so distinctly that it was ultimately separated and developed as a new, stand-alone series.
Named Unseen Flora (2023-2024), this new series offers a novel perspective on the world of fantastical botanical illustrations. Through a surreal representation of the natural kingdom, the work focuses on the visionary practice of four fictional British scientists and artists: Charlotte Bancroft, Beatrice Hastings, Edmund Thorne and Theodore Winslow.
The series, with a strong conceptual focus, has as its core the exploration of the idea of post-truth, contextualised in the contemporary digital age where the boundaries between real and virtual, natural and artificial, seem to become more and more blurred. Especially after the sudden influx of innovative technologies led by AI language models and image generation systems, the notion of truth has become even more arbitrary, always hidden behind a thin veil of plausibility and verisimilitude.
Unseen Flora stands at the interplay between these areas: the collection explores these concepts through four characters whose stories, while fictional, appear entirely believable to fresh, naive eyes. With deep influences from other research strands, such as Surreal Botany, and pioneering illustrators such as Ernst Haeckel and Luigi Serafini, this series of works offers an unprecedented glimpse into imaginary botanical dimensions. Developing through the stories, lives and creations of the four fictional botanists, it seeks to propose a reflection on the concept of truth and post-truth concerning the world of new technologies and, in particular, to AI language and image generators.
In this specific context, the perception of truth is shifting, seemingly aligning more closely with the notion of plausibility than concrete reality. For instance, post-truth indicates a deeply-rooted tendency to validate an assertion based on the public’s emotions and feelings, without any concrete analysis of the actual truthfulness of the facts being told. This tendency also taps into the concept of confirmation bias – the inclination to favour, believe, and recall information that reinforces one’s existing beliefs or values – which, if amplified on a global scale, may significantly influence and shape public opinion.
These concepts become even more relevant in the age of AI, now that models such as Chat GPT are regularly accessible: these systems leverage a precise human perception of the world that results from the initial, human-made training set. AI systems continuously generate plausible answers, post-truth realities that tap into what is commonly believed to be logical and credible, but not necessarily true. It is at this new intersection between these ever-evolving concepts of post-truth, plausibility and falsehood that we set the stage for Unseen Flora: a series that is founded on familiar but somehow uncanny botany illustrations, on plausible but fictional histories and personalities.
Creation Process
As mentioned, Unseen Flora involves a shift in Artificial Botany creative process: the past iterations of the work have always taken as a starting point authentic botanical illustrations, which were then reimagined through the lens of Generative Adversarial Networks (GANs). The central focus has always been exploring the transformative journey from natural forms to artificial ones, mirroring the organic morphing of real seedlings. Our tasks have always been confined to taking something already created, with a life of its own, and reinterpreting it through unprecedented techniques and looks. With Unseen Flora, however, we move backwards: instead of focusing on the outcomes of human-driven research, we delve into the lives of the botanists themselves, exploring their hypothetical perspectives on the natural world. An approach that allows us to generate unparalleled outputs that serve as starting points for further research and interpretations.
At the core of the process is a newly developed pipeline that integrates previously created animations using a StyleGAN model. A custom neural network extracts visual information from a single image and transfers it into Stable Diffusion, which then generates the next image in the sequence based on the extracted data. The result is a continuous generation of frames: StyleGAN produces an endless interpolation between forms, while Stable Diffusion enriches the sequence with novel elements and evolving features.
Artistic Influences
An essential element that shaped the visual research for the project was the exploration of works of illustrators and botanists who create unreal, imaginary worlds. Among those who inspired us most are the renowned Ernst Haeckel and Luigi Serafini’s Codex Seraphinianus. Haeckel (1834-1919) was a German biologist, naturalist, and artist best known for his intricate, artistic representations of marine life and other organisms. The Codex Seraphinianus, in contrast, is a whimsical book that has garnered increasing attention since its publication in 1981. It features a broad array of subjects, from biology and architecture to art and technology, all presented eccentrically and unconventionally, captivating readers with a unique view of the world.
Mimicry
Mimicry (2025) is a speculative exploration of evolution, adaptation and interspecies entanglement in an era of ecological disruption. The artwork imagines a world in which the natural world responds to anthropogenic change with surprising new forms of life – fantastical hybrids between insects and plants. Developed through a creative process mediated by artificial intelligence, Mimicry reflects on the resilience of biological systems and the unpredictable directions of evolutionary transformation.
While human activities rapidly shape the planet and threaten its biodiversity, evolutionary forces continue their slow processes of speciation. These forces are guided by the need to respond to new selective pressures faced by living organisms, generating adaptations in forms, colors, and interspecies interactions.
In a creative process analogous yet not identical to those at work in human artistic creativity, nature devises and tests surprising solutions to the new challenges introduced by human action.
Mimicry explores the boundaries and thresholds of fantastical species born from the hybridization of insects and plants, through a process mediated by artificial intelligence simulating possible natural evolutionary trajectories. These beings represent nature’s search for new survival strategies, such as exaptation – where forms originally intended for one function take on new and unexpected roles – and evolutionary mimicry, in which plants and insects mutually shape their morphological and behavioral traits.
Merging botanical and entomological aesthetics, Mimicry envisions unprecedented forms of coexistence and symbiosis among natural kingdoms considered distant, capable of adapting to a planet in transformation.
Concept
Evolution, with its random dynamics, generates a far greater variety of forms than those that actually manage to take hold, countless hybridizations that fail to pass the selective filters of survival. Humanity applies its own perceptual and evaluative criteria – shaped by aesthetic norms and utilitarian logic, choosing and privileging certain outcomes from nature’s vast field of potential.
Donna Haraway’s work offers a powerful lens for rethinking the categories through which we understand the world. In her seminal essay The Promises of Monsters (1992), Haraway describes monsters as entities that challenge the boundaries between organism and machine, animal and plant, subject and object. These hybrid, liminal figures exist to question cultural, scientific, and identity narratives, offering alternative visions of coexistence.
From the interplay between evolution and human filtering emerge these hybrid forms, which do not merely adapt to change; they inhabit zones of transition, becoming living metaphors of multispecies hybridization.
In the world of Mimicry, plants that resemble animals and insects that mimic leaves stage acts of evolutionary theater that echo Haraway’s notion of companion species: co-evolving beings shaped through deep, continuous relationships. Mimicry thus becomes a relational narrative, not just a survival strategy, but a choreography of presence, camouflage, and recognition – an act of becoming-with.
Within this framework, artistic creativity emerges as a radical tool for exploring the realm of the possible. It liberates the imagination from inherited aesthetic canons and strict notions of utility, allowing us to speculate, to fabricate monstrous forms rooted in real patterns of coexistence.
Research
The entities created in Mimicry respond to a context of profound ecological transformation, where the identity and distribution of Earth’s biomes are rapidly shifting due to global warming and human activities.
Recent projections based on the high-emission scenario (RCP 8.5), which predicts a global temperature increase of +3.5°C, indicate that by 2070, biome distributions will undergo drastic changes.
Tropical biomes are expected to expand, gradually replacing temperate forests and grasslands, while polar and alpine ecosystems will face severe contraction. At the same time, deserts and xeric shrublands are likely to spread, reshaping biodiversity patterns and ecological dynamics on a global scale. This accelerating transition highlights the urgent need to understand and anticipate the adaptive strategies of living organisms.[1]
In response to the changing landscape, the hybrid organisms of Mimicry evolve dynamically: part of their structure is optimized for biomes on the verge of disappearing, preserving traits suited to past ecological conditions, while another part adapts to the emerging conditions of newly forming biomes.
This dual adaptation allows them to bridge temporal and spatial ecological shifts, acting as transitional entities between ecosystems.
Through this process, they generate a living mosaic of organisms capable of crossing and interacting with multiple biomes, facilitating genetic and morphological exchanges that mirror the mechanisms of horizontal gene transfer – a process that allows organisms to acquire genetic traits from other species or kingdoms, thus speeding up adaptation. Although this phenomenon is known in microbes, its presence has recently been demonstrated in other kingdoms as well. In Mimicry, this transfer enables plant-insect hybrids to continuously transform, absorbing and incorporating characteristics from both kingdoms.
Creation Process
The botanical structures of Mimicry are generated in continuity with Unseen Flora, establishing a visual link between the two projects. Using text-to-image models, these plants are created based on real clades within the Plantae kingdom, with prompts describing their morphological traits. The insectoid elements are then integrated through image-to-image processing, working with an archive of entomological illustrations to merge plant structures with insect-like forms. Diffusion Models have also been used to synthesize images by gradually transforming patterns of noise into coherent visual forms, guided by input data such as text or reference images.
Figure 14: Horizontal gene transfers
The initial botanical forms are generated using text-to-image models, with prompts describing real plant clades and their morphological features. These images establish the base material for further transformation.
The creation of hybrid insectoid structures involves two image-to-image stages. First, the botanical images are reinterpreted using a curated archive of entomological illustrations, allowing plant forms to take on insect-like qualities. Then, additional transformations refine the hybrid through a second image-to-image pass, blending visual traits from both kingdoms. Throughout the process, text prompts also accompany the insect component, guiding the evolution of specific features such as segmented limbs, wings, or exoskeletal textures.
This layered approach, merging semantic control with visual references, enables a continuous transformation between botanical and insectoid identities.
Figure 15: Diffusion model for Mimicry
References
La botanica parallela – Leo Leonni, 1976
Codex Seraphinianus – Luigi Serafini, 1981
The Saturated Self: Dilemmas of Identity in Contemporary Life – Kenneth J. Gregen, 1991
The genetic basis of a plant–insect coevolutionary key innovation – Christopher W. Wheat, Heiko Vogel, Ute Wittstock, Michael F. Braby, Dessie Underwood, and Thomas Mitchell-Olds, 2007
Brilliant Green – Stefano Mancuso, 2015
The Art and Science of Ernst Haeckel – Julia Voss and Rainer Willmann, 2016
The Promises of Monsters: A Regenerative Politics for Inappropriate/d Others – Donna Haraway, 2019
Intelligenza Artificiale – Luciano Floridi & Federico Cabitza, 2021
Trait-based projections of climate change effects on global biome distributions – Coline C.F. Boonman, Mark A.J. Huijbregts, Ana Benítez-López, Aafke M. Schipper, Wilfried Thuiller, Luca Santini, 2021
Genomics reveals widespread hybridization across insects with ramifications for species boundaries and invasive species – Michael San Jose, Camiel Doorenweerd and Daniel Rubinoff, 2023
Deep reticulation: the long legacy of hybridization in vascular plant evolution – Gregory W. Stull, Kasey K. Pham3 , Pamela S. Soltis and Douglas E. Soltis, 2023
Entomology Redefined Current Trends and Future Directions – Abdul Rahaman M Nadaf, Venukumar S, S. Vinoth Kumar, Thakar Pratikkumar, Sangavi R, 2023
Long‑term data in agricultural landscapes indicate that insect decline promotes pests well adapted to environmental changes – Tim M. Ziesche, Frank Ordon, Edgar Schliephake, Torsten Will, 2024
The angiosperm radiation played a dual role in the diversification of insects and insect pollinators – David Peris & Fabien L. Condamine, 2024
Weather explains the decline and rise of insect biomass over 34 years – Jörg Müller1, Torsten Hothorn, Ye Yuan, Sebastian Seibold, Oliver Mitesser, Julia Rothacher, 2024
Footnotes
Source: Trait-based projections of climate change effects on global biome distributions (2021). www.researchgate.net/publication/356065146_Trait-based_projections_of_climate_change_effects_on_global_biome_distributions ↑
Source: Trait-based projections of climate change effects on global biome distributions (2021). www.researchgate.net/publication/356065146_Trait-based_projections_of_climate_change_effects_on_global_biome_distributions ↑
- This article was published under the group designation fuse* as an author pseudonym, as the text represents the joint work of the artist collective fuse*, who speak here with a unified voice. ↩︎