Participatory design and Artificial Intelligence

3 June 2025. Published by Benoît Labourdette.

6 min

When participative creation and AI meet, four modalities of use and two artistic schools intertwine. A practical guide to navigating between innovation and mastery.

Two approaches to participative creation

Participative creation is a multifaceted field of artistic action, often led by an artist or artistic team. To simplify, we can distinguish two schools or approaches: the participative approach and the cooperative approach.

The participative approach:
In the participative approach, participants take on the role of performers serving the artistic project conceived by the artist. Even when the work—whether theatrical, literary, choreographic, or musical—is developed from participants’ contributions, the artist maintains final control over the result. This process can prove deeply enriching, provided that the rules are clearly established and everyone understands the decision-making power that the artist maintains throughout the process.

The cooperative approach:
The cooperative approach operates from a very different dynamic: the artist designs encounter protocols, prioritizes the interaction process, and accepts subjecting their own ideas to the test of collective contributions. This method can completely transform the distribution of roles and responsibilities initially envisioned. In a theater production, for example, the artist who is usually the director might become an actor, while another group member takes on the directing role. Moreover, certain creations can emerge even in the absence of the initiating artist, thanks to the conditions and spaces for encounter they have established. Theater director Pippo Delbono works extensively in this manner, which is why in his performances, one senses a form of absolute rightness in the actors’ presence on stage and their interaction with theatrical images, because they have often formed them together, in this shared universe where they have their full place of invention and expression. This is also how I worked on the staging of the short show Bonjour ma belle (video recording) 10 years ago for students at the Fémis and CNSAD, leaving the students alone for a while, in a certain spirit, so that their perceptions of the future, which was the subject of the show, could emerge on their own.

The risk-taking is much greater because the result is much less predictable than in the other process, and it is a longer, more demanding process that requires, contrary to what one might think, much more work and much more finesse in balance and attention to all the bonds that form the creative team assembled for the project.

Four modalities of artificial intelligence use

Artificial intelligence, in the way it is employed in artistic creation, also has distinct modes of use, four in my opinion:

1. The augmentative modality
In the augmentative modality, artificial intelligence is used as a tool to save time, create syntheses, for example from the transcription of a conversation between people. We can even ask artificial intelligence to write a theater scene that incorporates ideas shared in the conversation. Yes, there is a small creative component, but we’re more dealing with a “ghostwriter,” as it was called (what an ugly word), that is, an extremely competent person at writing, who writes for others without signing the books, but who has no ego, no identity, like AIs...

And artificial intelligence doesn’t just write. In the augmentative modality, we can produce images or music, for example.

2. The creative modality
This approach is the one that interests me most. We will instruct the machine through different entry points to invent, to work with chance. Here, it’s up to us to establish protocols—I will propose some—so that artificial intelligence can truly be a collaborator. It’s not about making syntheses or doing “in the style of,” but rather that it feeds on what occurs to be in creative interaction.

Several setups can be implemented:

Direct oral exchange: A phone equipped with a conversational assistant with voice recognition actively participates in creative discussions.
Visual interaction: The AI regularly receives photographs during work sessions, documents, images, gestures, and generates creative proposals in response.
Inter-AI dialogue: A particularly promising setup where two distinct artificial intelligences (ChatGPT and Claude, for example) dialogue with each other based on the same source documents. They co-construct, question each other, debate, or are even invited to contradict each other, to develop together, for example, the text of a scene. If we put two artificial intelligences in discussion with each other, we’re not in simple augmentation, there will really be a creativity specific to the confrontation between the reasoning modalities of these machines, which, due to the interaction, will multiply creative capacities that are necessarily quite limited in a given model, but which, through the confrontation of two different models, can, in my opinion, open doors to very important narrative discoveries, for example.
AI mediator: The AI can also serve as an interlocutor within a human collective, participating in conversations and providing structure or questioning. If humans wish to freely explore bold ideas, the AI can progressively structure their proposals through its questions, with text displaying in real time on screen during exchanges. The AI’s listening capacity represents a major asset. Where humans struggle to maintain genuine listening, project their own ideas, quickly tire from this energy-consuming exercise, the machine remains intrinsically listening, without fatigue. This quality makes it a valuable partner for fostering creative cooperation between people.

3. The subject modality
Artificial intelligence here becomes the central theme of the work, the subject of the performance, film, or poetry collection. This approach has existed for a long time in theatrical, literary, and cinematic science fiction, well before the democratization of conversational AIs in November 2022. The film project A.I. Artificial Intelligence by Stanley Kubrick, ultimately directed by Steven Spielberg years after Kubrick’s passing, constitutes an emblematic example.

In the context of participative creation, this modality presents particular challenges. The attention given to process as much as result, an essential characteristic of this approach where the result often benefits from telling its own process rather than freezing into a finished aesthetic, complicates the integration of AI as a subject.

One possibility would be to organize human co-writing on the theme of AI, for example around the idea of an AI psychologist. Participants would exchange on this subject while progressively testing, on their own, concrete experiences with AI-therapists, then sharing their discoveries to enrich the theatrical project. This approach seems ultimately more promising than I initially envisioned.

4. The interactive modality
Particularly suited to live performance, this modality exploits scenic or participative automation devices. For example, a performance where the audience would ask questions via smartphone, receiving responses from an AI based on the ongoing stage action.

A more elaborate device could comprise three spaces: a zone for spectators and two scenic zones where audience volunteers, briefly prepared, would become actors. In one, they would perform a short scene repeating with improvised variations. In the other, participants would dialogue via their phones with the AI that would reveal, for example, the unconscious thoughts of the observed characters. The phones would be filmed live, with responses displaying on a large screen, bringing completely unpredictable elements to the performance.

This format could work in 15-minute cycles, with rotation of participants between stage and audience, creating continuous interaction where the AI would play an essential autonomous role in the artistic device. This dimension of autonomy constitutes precisely the most fascinating aspect of contemporary AIs.

Crossing approaches: stakes and precautions

The intersection between the two modalities of participative creation and the four uses of AI offers multiple possibilities, but requires great clarity in the chosen protocol and in what the artistic team is ready to welcome.

Indeed, a cooperative creation project integrating a creative AI can take completely unexpected directions. This openness requires a work framework, time, and technical means capable of absorbing significant changes. When these conditions are not met, it may be wiser to opt for classic participative creation assisted by augmentative AI, or to choose AI as a subject within a known cooperative framework, to limit destabilization risks.

The main danger in any participative project occurs when the “supporting structure” (in the symbolic sense, it’s generally the artist’s role), because one is always needed, even if the bearer is not necessarily a chief, feels overwhelmed by events. This symbolic structure, necessary for gathering participants and compatible with cooperation, can then tense up, become rigid, and lead the project to the opposite of its initial ambitions.

One must be very careful in this type of project, carefully measure one’s ambition and means. Often we want to do a lot, we want to do the brilliant project where everything is completely participative, artificial intelligences are everywhere, etc. Be careful, it might be better to do a little less, but be in great accuracy, in great respect for everyone, and not find ourselves overwhelmed, having to ultimately take back power so that things can be finalized and arrive at an effect opposite to what we initially wanted.

This can happen. And it’s precisely for this reason that I share this methodological reflection. It’s to set milestones and better understand where we’re stepping and what we’re capable of welcoming, and also what interests us in doing at this intersection between participative creation and artificial intelligence.