©ABK
On the occasion of the Summit for Action on AI (SAIA) in France, and in connection with her artistic and academic research, the artist-doctoral student Alexandra Boucherifi-Kornmann organized, with the support of the artist-researcher Olga Kisseleva, two seminars on AI and creation in November 2024 and February 2025, at Paris 1 Panthéon Sorbonne and at the Maison de la Conversation, in Paris. The events, entitled respectively AI and creation: the best of codes? and AI and creation: it decodes completely!, allowed to cross-reference knowledge, to question the challenges of creative AI and to conduct a critical reflection on these issues on ethical and aesthetic levels.
These days dedicated to AI and creation brought together numerous speakers, including Alexandre Gefen*, who introduced the discussion with a long historical overview of artificial intelligence. Researchers, artists, experts, and lawyers then explored reflections on the interactions between artificial intelligence and creative processes in various fields such as the visual arts, cinema, writing, and music. These events contribute to the development of an inventory of the impact of AI on contemporary artistic practices, highlighting the way in which it accompanies and transforms our uses, whether visual or everyday. Discussions also focused on the philosophical, ethical, and aesthetic issues raised by these new practices. In this new era where machines seem to be becoming co-creators, the use of AI encourages us to question, in particular, the notion of authorship and the future of human creativity. Thanks to the speakers' research, and to discussions between researchers and the public, these days opened up perspectives for imagining the future of creative practices in a post-digital world. Here's a glimpse of the presentations presented by the researchers.
Painting the humaniplant using the brush and AI by ABK
Alexandra Boucherifi-Kornmann is a co-supervised doctoral student (Bordeaux Montaigne and Paris 1 Panthéon Sorbonne universities). Her topic is augmented painting or augmentation, a new pictorial movement that includes AI.
With Painting the Humaniplant Using a Brush and AI, ABK proposes the creation of a new kind of creature. At the origin of the concept of the Humaniplant, a hybrid figure combining humans and plants, ABK questions our relationship with living things in a changing world. The integration of AI into her creative process allows for an innovative exploration of this hybridization, with AI generating forms and visual evolutions combining technology and pigment painting. The combined use of a traditional brush and artificial intelligence offers a new dimension to this visual exploration, which she puts into perspective with Margaret Boden's approach to transformational creativity. Through these plastic experiments, the artist raises questions about transformation, identity, and the blurred boundaries between species and even between kingdoms. The Humaniplant becomes a metaphor for our time, marked by the rapid evolution of science, technology, and the transformation of the environment. Through his work, the artist invites us to consider speculative futures where biological boundaries are redefined in the light of biotech, as well as to rethink our relationship with nature, to better reconnect with it. Series initiated in 2015, of which here are some examples until 2019: click .
©ABK
Real Intelligence by Olivier Auber
Leo Apostel Center for Interdisciplinary Studies (CLEA) at the Vrije Universiteit Brussel (VUB), artist.
Under the title Intelligence aRéelle, Olivier Auber presented the foundations of a novel artistic and scientific experiment that he will launch shortly. Artificial agents of the LLM type will be invited to interact freely in an AI network according to rules borrowed from the Poetic Generator (PG), a historic digital artwork from 1986 that was a precursor to social networks. Unlike chess or the game of Go, where the goal is to "win" and where AIs have surpassed humans, there is no predefined goal in the GP, nor any notion of winner or loser. Researchers will observe how these artificial intelligences, each drawing in response to the drawings of others, will potentially develop their own norms, culture, and temporality—a phenomenon dubbed “aReal Intelligence.” This interdisciplinary research, open to all, will allow researchers and enthusiasts to deploy their favorite LLMs and analyze their interactions. Beyond the artistic experience, this project questions the fundamental mechanisms by which shared realities emerge, whether human or artificial. While waiting to see AIs flock to the GP, we can play with its human version: clicking .
©Olivier Auber
Creative Neuro Divergence, the Atypical Model of Generative AI by Maurice Benayoun
Chair Professor Nanjing University, School of Art and School of Architecture and Urbanism, Head of the Art and Architecture Intelligence Lab, artist.
Creator of the title of the seminar AI and creation: It totally decodes!, Maurice Benayoun discussed creative neuro-divergence, the atypical model of Generative AI. "When we talk about artificial intelligence, we can have the impression of dealing with a unique mechanism or machinism. However, for several decades we have begun to take into account what we call neuro-divergence: a way of thinking that does not correspond to the standard model of so-called neurotypical intelligence, and which presents specificities often associated with cognitive dysfunctions such as schizophrenia, ADHD or autism spectrum disorders. Other studies note the association that could be established between creativity and neuro-divergence; symptoms that are found in individuals known to be creative and drivers of innovation. Latent space management as practiced by generative AI could use similar mechanisms based on associations and cognitive chains that could bring out results that have a certain level of similarity with creative productions while producing a satisfaction/seduction index that would facilitate their reception and adoption. To summarize the application of the neurodivergent model to the generation of text, music or images, one would have to imagine a flow of thought that would follow, like water flowing down a mountain, its line of greatest slope. This mechanism induces creative processes that draw their singularity from the activation of concepts specific to AI, giving rise to complex and relevant forms for their capacity to respond beyond the request while providing elements of questioning such as: iterative curation, augmented serendipity, maieutic suggestion engine, artificial intentionality… These conceptual bricks can contribute in their dynamic form to the development of art-subject as a surpassing of art-object consecrated by millennia. Maurice Benayoun's communication was thought out from examples taken from his practice freely combining Virtual Reality, interactivity, immersion, electroencephalography, AI, robotics, blockchain, which find their meaning when they are applied to societal issues such as finance, ethics, language, poetry and politics… Website of the artist-researcher, click .
The Struggle of Natural Intelligence with Artificial Intelligence, Maurice Benayoun + AI, 2023.
Machine Introspection: Creative Fictions of the Artificial Unconscious by Robin Champenois
Doctor of Artificial Intelligence and Artistic Creation, SACRe/ENS, artist
Under the title Machinic Introspection: Creative Fictions of the Artificial Unconscious, Robin Champenois explores how artificial intelligence, beyond its role as a generative tool, becomes a true creative partner, carrying a kind of its own subjectivity and a singular imagination. Beyond its simple generative capacities, AI represents, in the artistic field, a very particular tool: it can be endowed with a strong personality; to use it, it is necessary to engage in a form of dialogue. Indeed, as complex learning machines, neural networks do not behave like logical and rational processes: their structure and skills more closely resemble our intuitive and rapid analytical capacities. A kind of Artificial Unconscious, whose sophisticated biases reflect our own. How can an artistic approach be articulated with this Artificial Unconscious, draw its contours, reveal its rough edges? What visual language can emerge from friction with the machine? What do these algorithms, trained on our data, say about us? How can works of art help us encounter the radical otherness of the machine, and turn our gaze back on our own (non-)human relationships? Where commercial actors fear the deviations of their algorithms, and strive to "align" their models as best as possible, artistic research can nestle precisely in the search for these excesses, these unforeseen events. These "Happy Accidents." These evocative, troubling, revealing oddities. And lead us to encounter the unreasonable fantasies of the machine and its creators.
ART*ificielles: Consciousness and Creativity by Aida Elamrani
IPEM Researcher, Fac. of Arts and Philosophy, Ghent University
ART*ificielles: Consciousness and Creativity explores the blurred line between artificial intelligence and subjectivity, questioning a machine's ability to transcend its status as a tool to access a form of autonomous thought. Is the mind reduced to the mechanical processing of information by our bodies? In this case, a computer could truly have mental states. This concern haunts researchers in the science of consciousness, a highly interdisciplinary field of study that has been formally established for some thirty years. The most elusive aspect of our mental life is phenomenal experience: an ultimately immaterial corner of our minds, which eludes any attempt at functional characterization. To grasp this, some researchers have proposed approaching it through the prism of creativity. An AI would be truly creative to the extent that its actions would no longer be explainable from its code and data. Here, "creativity" does not refer to a conceptual field, but to the very origin of thought. Animated by a breath of life, by a psyche, the machine would then join the realm of the conscious.
Questioning uses with Daphné Greiner
PhD student at the Panthéon Sorbonne doctoral school of management, specializing in the use of humor by virtual agents
Artificial intelligence (AI) is profoundly transforming our relationship with creation. By making it possible to compose music, generate images, or write texts, it is establishing itself as a creative partner that questions the very definition of the artistic act: are we truly inventing with AI, or is it simply rearranging what already exists? The example of humorous AI, such as on the Beevi.fr website, capable of generating or analyzing humor, illustrates this ability to push the boundaries of what is possible while confronting us with a new idea of creation, where the tool sometimes becomes the object itself. However, these innovations pose major challenges. The use of these technologies often remains intuitive, like smartphones that we master without knowing their inner workings. In a context where regulatory safeguards are still lacking, it is up to everyone to question their uses and equip themselves with mechanisms to regulate these tools. Curiosity and vigilance are required to prevent these allies from becoming alienating instruments. Creating with AI therefore means exploring a delicate balance between assistance and substitution. These tools, while multiplying our possibilities, remind us that their true value depends on the human intentions behind them. More than ever, it's about combining technological innovation and ethical reflection to make AI a lever for emancipation and not a hindrance to our humanity.
From Swarm Intelligence to Artificial Intelligence and Back Again by Olga Kisseleva
Lecturer HDR Art&Science in Arts and Art Sciences at the University of Paris 1 Panthéon-Sorbonne, and artist
Olga Kisseleva presented a talk entitled From Swarm Intelligence to Artificial Intelligence, and Vice Versa. With a background in mathematics and art, Olga Kisseleva has been exploring AI since the early 2000s. Examples include Power Struggle (2011) and EDEN (an art project started in 2012). Her works address the links between AI and the intelligence of nature, such as her recent project AntiCorps (2020-2021) and the series Cities Live Like Trees (2022). Her presentation drew on this extensive experience, as well as her recent involvement with the Sorbonne's AI Observatory and the CNRS's CulturIA project. Within this theme, she addressed the notion of intelligence and the role of art as a mediator for deploying and inventing new objects and new languages. She also raised the question of how AI can be improved by natural intelligence and, conversely, how AI can contribute to repairing our world, from an ecological perspective. The artist emphasizes the danger of capitalism co-opting certain biotechnological research and sees her artistic approach as a way of diverting AI, of proposing other forms of networks and connections.
AntiBody, Olga Kisseleva, 2020-2021.
The State of AI Consistency and Controllability in Cinema by Louis Laborelli
Researcher at INA
Artificial intelligence (AI) is advancing rapidly in video creation, and the question of its ability to generate a complete film from a single prompt remains open. One prediction market estimates this possibility at 45% by early 2028. By late 2024, the field of AI video generation is very active, distinguishing three levels of tool availability: company announcements on poorly available tools like Sora, opaque and paid solutions, scientific articles without implementations, and fully transparent open source resources. Quality criteria include the consistency, controllability, and graphic quality of the sequences. Inspired by traditional cinema, the goal is to achieve precise control over framing, camera movements, actions, dialogue, and lighting via interfaces or text instructions. Tools already allow the generation of videos from text descriptions (prompts) or images, with combined approaches. Companies like Kling, Minimax, and Runway ML are innovating in this area, each with its own strengths: Kling offers precise control of paths with motion brushes, Minimax better follows text instructions, and Runway ML offers tools for controlling camera movements. Despite these advances, challenges remain, such as controlling actors' gaze, compositing with real-world images from moving cameras, and ensuring lighting consistency between real and generated images. A more advanced 3D interface could improve control of actor and camera movements, replacing current purely 2D approaches.
Post-Realism in the Age of Generative AI by YAK (Yacine Aït Kaci)
Creative artist of ELYX (digital ambassador of the United Nations) and co-founder of the ELYX Foundation
YAK's intervention, Post-realism in the Age of Generative AI, presented a reflection-transformation of our artistic and cultural references in the era of creative artificial intelligence, where image and narrative are emancipated from any material anchoring to recompose themselves in a flow of infinite self-references. We are the generation that will have experienced the closure of the first dataset of artificial intelligence, better known as the history of Art and Civilizations. 73,000 years separate the first drawing of humanity and the first images created by generative AI. This first generation represents only a tiny part of the data that ends up self-referencing. In a race to imitate reality, the truth has been lost somewhere between intersubjective realities and lies replicated on an industrial scale. The post-truth era contributes to the shattering of a world based on the sharing of a common experience. It is in this context that the proposition of Post-Realism opens up for art. "In a post-truth era, the only thing we can truly believe is what does not claim to be real." This is the starting point of the universe of Archipel, imagined and created by Yacine AIT KACI, for a series of exhibitions that began last January at Polaris Centre d'Art and will be held in Nice on the occasion of the UNOC, the United Nations Conference on the Ocean.
©YAK
*Research Director at the CNRS, main leader of the ANR CulturIA project, CNRS research director within the Theory and History of Modern Arts and Literatures unit.
Nicolas Sarazin
AI Suggestions