AI in Service of Democratic Transformation of Cultural Institutions

Artificial intelligence is most often deployed to optimize production chains, for example through their digitization to create a digital twin in order to find ways to reduce costs and increase efficiency. In the subsidized cultural field, the objectives are very different from those of commerce. The challenge and potential of AI does not consist, in my view, in automating tasks for increased performance; AI can serve to reveal the real practices of audiences and agents, with the aim of promoting a deep cultural democracy, respectful of everyone’s rights and dignity. I propose a seven-step methodology, derived from industrial process management systems, but reinvented to prioritize openness, cooperation and the discovery of the unexpected, and to place humans and territory at the center. The goal is to avoid the traps of rigid optimization, by favoring collective dialogues and an ethics centered on authentic relationships, so that cultural venues can, also thanks to AI, rebuild themselves even better as true co-constructed commons.

The paradigm shift: from optimization to revelation

Supply chain optimization through massive data analysis to minimize losses and accelerate flows, focusing on metrics like costs and speed, is a quest for efficiency that may be useful in industry, but it risks dehumanizing processes if applied as such to the cultural sector; we must guard against simplistic translations. In a cultural institution, the challenge is the reverse: AI can be a very powerful tool to better reveal real practices, with the aim of serving cultural democracy and honoring the dignity of each person, by allowing deeper exploration of how the venue interacts with its territory and its audiences, including those who are excluded from it.

The traditional approach of surveys and optimization processes locks us into predefined frameworks that stifle possibilities for cultural and social innovation, amplifying the biases inherent in AI such as cultural stereotypes or the standardization of perspectives. Instead, I advocate for an exploratory stance where AI helps welcome surprises, by asking non-directive questions to discover invisible patterns and reveal symbolic barriers to cultural participation.

This reversal integrates a relational ethics, where responsibility is shared in a network of human actors, to avoid unpredictable impacts on human relationships. It involves cultivating shared resilience in the face of uncertainty, distinguishing authentic empathy from technological or statistical simulations, for a transformation that enriches living together rather than rationalizing it. Often, “audience surveys,” for example, done with or without AI, are confined to pseudo-rationalization, completely foreign to the real movements of enrichment between human beings, which is at the heart of cultural democracy, which is in my view the path for tomorrow’s development of subsidized culture.

Methodological approach adapted to the cultural sector

Here is a seven-step approach to implement these transformative methods through AI. I hope it is inspiring, in its overall logic or in its details.

Step 1: Form a reflection collective (not “appoint a manager”)
Instead of designating a manager who imposes a top-down vision, let’s create a cross-cutting collective including reception staff, artists, programmers, audience representatives such as associations and regular users, non-audiences from the territory, technicians and administrators. Animation roles rotate, and decisions are made by consent to foster authentic cooperation.

This collective collectively documents exchanges, capturing without filter to preserve the diversity of voices. This avoids rigid hierarchies, integrating varied perspectives, to co-construct an emergent understanding, aligned with participatory governance (to resist technocratic approaches). This group becomes a secure space where tensions enrich dialogue, cultivating irreplaceable empathy and intergenerational transmission to invent cultural mutations together.

And AI allows synthesizing these exchanges from their traces (written, recorded, photographed), to help create narratives and legitimize them, as their analysis will show their relevance.

Step 2: Listen to the territory (not “identify beneficiaries”)
Rather than defining “clients” or “targets” according to marketing criteria, organize active listening sessions in and around the venue: spontaneous conversations recorded with dictaphones, dialogues with passersby who avoid entering, exchanges with local merchants and associations.

Documentation is open: audio recordings, photos of spontaneous uses, observation notebooks, stories of “non-encounters” explaining absences. This reveals invisible barriers and hidden sociability networks, without a predefined grid to let the unexpected emerge.

This listening favors cultural diversity, integrating non-Western or marginalized perspectives, and protects sensitive data for respectful ethics, transforming the institution into a living reflection of its territory. AI will enable producing structured syntheses of all these informal exchanges, from their traces, and will thus be able to reveal paths for creation, interaction and cooperation that would never have appeared otherwise.

Step 3: Map real practices (not “the current process”)
Instead of mapping formal processes like reception or ticketing, AI will be able to analyze conversations to reveal the real journeys of audiences, diverted uses of spaces and usage temporalities, such as creative off-peak hours or social micro-rituals.

Visualizing human flows includes not only those entering, but those who pass by without entering, identifying surprising patterns without a priori categorization. This confronts AI biases by favoring ethical traceability, with open questions like “Which voices are missing?”

This mapping becomes a discovery tool, linking heterogeneous data for nuanced understanding, ready to integrate the ecological impact often hidden in digital tools.

Step 4: Define dignity indicators (not “performance” ones)
Rather than occupancy rates or revenues, measure the diversity of voices heard, the quality of encounters through individual stories, the sense of belonging, spontaneous initiatives, personal transformations and new social links created.

These human indicators place respect for singularities above efficiency, evaluating openness to the unexpected and rediscovered meaning. They integrate vigilance against discrimination, training on gender or ethnic biases, which are often amplified by AI.

Thus, evaluation becomes democratic, valuing resilience and collaboration as keys to an inclusive and adaptable culture. AI will enable creating these evaluations with dignity indicators (which would be too biased if produced by human agents).

Step 5: Co-construct with creative AI (not “seek facilitating tools”)
Instead of automating ticketing for example, use AI for creative dialogues: have two people debate “What is a democratic cultural venue?” to bring out the unexpected, or as a mediator by synthesizing exchanges in real time and asking open questions about the unspoken.

AI analysis should be prompted in the direction of AI creativity from this data: “What surprising models emerge?” or “What contradictions appear?”, for example, avoiding orienting the prompts. This transforms AI into a discovery partner, as it will allow linking heterogeneities to reveal the invisible.

This co-construction cultivates vigilant optimism, distinguishing simulation from authenticity, and favoring critical experiments that question established cultural norms.

Step 6: Experiment with transformations (not “rethink the process”)
Test new co-created hospitality rituals, dare to create spaces of freedom for audience appropriation, programming emerging from revealed desires, reversed mediations where audiences transmit their knowledge, or alternative temporalities like night openings, cultural breakfasts or any other atypical and inspiring experiences that came from dialogues.

These small-scale experiments should be documented with AI help, which will be able to help measure their impacts continuously, with more ease for participatory adjustments, thanks to syntheses made very quickly by AI, always placing shared human evaluation as priority. AI will allow easily integrating the stories of this irreplaceable materiality, valuing processes rather than results, to preserve embodied experience.

Step 7: Continuous and open evaluation (not “control”)
Practice monthly circles open to all, with sharing stories of transformations rather than statistics, documenting failures as learning, permanent adjustments and celebrating unexpected discoveries. AI will allow very easily giving meaning to failures for example, and thus daring to share them more and advance in sincerity and sensitive openness to each other.

This democratic evaluation through AI measures respect for dignity, questioning meaning and integrating the societal dimension, to resist standardization. It also allows, through its relative neutrality, to maintain ecological and ethical vigilance, indicating discriminatory uses.

Thus, it enables piloting nuanced transformation, preserving authentic relationships and cultivating resilience in the face of profound social changes linked to technologies and political mutations.

Benefits specific to the cultural sector

Beyond efficiency
This methodology restores the legitimacy of institutions as commons, favoring social innovation where new forms of living together emerge. It embodies cultural democracy by making inhabitants actors, clarifying the public service mission beyond numbers.

Venues reconnect to their territory, revealing and integrating marginalized voices for a living and inclusive culture, resistant to technological standardization.

Finally, it liberates collective creativity, welcoming the unexpected as a source of mutual enrichment and shared meaning.

Impacts on teams
Teams rediscover motivation through renewed meaning, and can develop skills in facilitation, active listening and reversed mediation. Pride emerges from truly serving the territory, with this creativity liberated by openness to external perspectives.

Professional resilience increases, learning to navigate uncertainties with empathy and collaboration, valuing authentic relationships, even if they don’t seem “efficient” at first glance (it’s AI that will give them this status).

This internal transformation reinforces versatility, valuing intergenerational transmission and adaptation, for more united and innovative teams.

Points of vigilance and ethics

Resist drift
Do not instrumentalize AI to monitor or control, protecting participant anonymity and guaranteeing the right to withdraw at any time (respect GDPR, obviously). Avoid the race for numbers, even with dignity indicators, to prevent dehumanization and performative obsession.

Maintain vigilance against amplified biases, training agents on discrimination issues. Prioritize sober technical solutions to minimize ecological impact, collectively questioning the relevance of tools. Here too, don’t try to go too fast (being falsely “efficient”), but instead consider that confronting viewpoints with respect for others will open the most relevant path (and this thanks to AI which will take care of writing syntheses, assessments and analysis elements).

This “ethics by design” integrates reflective moments, distinguishing real empathy from illusions, for critical and humanist appropriation.

Maintain the democratic course
Ensure shared power without imposed decisions, with total transparency of documents, which must be accessible to all, at the end of meeting moments, and in a technically very simple way (a QR Code, to access a “drive” without need for registration). Cultivate hospitality to welcome the unexpected as an opportunity.

Mixed committees evaluate human impacts regularly, democratizing debates through citizen consultations, which will seek the voice of young people and marginalized persons. It’s about questioning the societal model, preserving authenticity and diversity.

Thus, governance remains inclusive, exploring liberated time for well-being and engagement for transformation guided by humanist values, implemented in the interaction modalities in proposed encounters, whether formal or informal. For example, the QR code can be transmitted as much to someone who participated in a long meeting or shared creation process, as to someone with whom we discussed for 5 minutes on a sidewalk or in a market.

A gentle but profound revolution

This approach, which marries AI uses and collective intelligence techniques, allows I hope cultural institutions to rediscover themselves through their territory’s gaze, to transform without violence, to reconnect to their democratic mission and to reinvent themselves continuously and facilitated. The essence of such a continuous approach facilitated by AI is to discover the unanticipated, gradually revealing what can be a shared, co-constructed and living culture. AI is no longer an efficiency tool but a discovery companion, in service of a humanist and democratic vision.

To deepen this, I cross this reflection with Olivier Hamant’s concept of robustness, in his book Antidote to the cult of performance (2023). He opposes the robustness of the living to the fantasy of performance:

“To inhabit this world without questioning performance would be madness.”

“Robustness is what allows maintaining the system stable in the face of fluctuations.”

“War is therefore both a product and a cause of performance gains, in an endless gear.”
“When a measure becomes a target, it ceases to be reliable.”

“Growth gives the illusion of abundance, while it creates scarcity.”

“The living is neither effective (it has no objective) nor efficient (it wastes enormous energy and resources). It mainly hosts a myriad of counter-performances at all scales, from molecule to ecosystem.”

“We will now have to live in a fluctuating world, that is, invent the civilization of robustness, against performance.”

“Performance only serves individual comfort by excluding others, human and non-human alike, robustness makes links the lever of balance and survival of the group.”

These ideas align with my plea for robust cultural institutions, integrating redundancies, chances and heterogeneities to face turbulence, rather than fragile optimization. By adopting this robustness, cultural venues become adaptable and human, ready for a fluctuating world, reconciling discovery and dignity.