IA Right
Individual Sovereignty over Intangible Cultural Heritage as a Foundation for an Ethics of Artificial Intelligence
Reflection based on the UNESCO Convention of 17 October 2003
Abstract
The Convention for the Safeguarding of the Intangible Cultural Heritage, adopted by UNESCO on 17 October 2003 and ratified by 180 States, establishes a fundamental yet often overlooked principle: the sovereignty of the individual in recognizing their own intangible cultural heritage (ICH). This paper explores the implications of this principle for reflection on artificial intelligence, articulating the distinctions between knowledge and know-how, between open and free, and between explicit and tacit knowledge. It proposes that this international legal framework, which places the individual at the heart of heritage recognition, offers a model for thinking about an ethics of AI that respects individual and collective sovereignties in the face of data extraction and aggregation logics.
1. Introduction: a widely overlooked legal framework with profound implications
The UNESCO Convention of 17 October 2003 for the Safeguarding of the Intangible Cultural Heritage represents one of the most widely adopted international legal instruments, with 180 States parties. Yet its implications for emerging technologies, particularly artificial intelligence, remain largely unexplored.
Article 2 of the Convention defines intangible cultural heritage as:
"the practices, representations, expressions, knowledge, skills – as well as the instruments, objects, artefacts and cultural spaces associated therewith – that communities, groups and, in some cases, individuals recognize as part of their cultural heritage."
The phrase "in some cases, individuals" is of capital importance. It establishes that recognition by the individual is legitimate and sovereign. A person may consider that a practice, know-how or expression forms part of their cultural heritage, independently of any prior collective validation. This provision anchors in international law the individual dimension of intangible heritage.
2. Conceptual clarifications: knowledge, know-how, open, free
To fully grasp the implications of this principle for AI, several conceptual distinctions are necessary.
2.1 Knowledge vs Know-how
The English language operates a distinction that French tends to blur:
| Concept | Term | Nature | Transmission |
|---|---|---|---|
| Knowledge | Knowledge | Theoretical, declarative (knowing that) | Through reading, study |
| Know-how | Know-how / Skill | Practical, procedural (knowing how) | Through practice, imitation |
This distinction overlaps with that introduced by Michael Polanyi, between explicit knowledge (codifiable, transmissible through writing) and tacit knowledge (embodied, difficult to formalize). Polanyi's famous aphorism captures this difference: "We can know more than we can tell."
2.2 Open vs Free
A second distinction, inherited from the free software movement, is crucial:
| Characteristic | "Open" Knowledge | "Free" Knowledge |
|---|---|---|
| Definition | Accessible and reusable without major barriers | Guarantees the four fundamental freedoms (use, study, modification, sharing) |
| Gratuity | Yes | Yes |
| Modifications | Possible, but conditions vary | Fundamental |
| Licenses | CC BY, CC BY-SA, CC0, etc. (broad spectrum) | Narrower spectrum (CC BY, CC0, debate on CC BY-SA) |
"Free" knowledge constitutes a more restrictive subset of "open" knowledge: it requires not only access but the effective possibility of modification and redistribution. This distinction is essential for understanding what AI can do with the data it absorbs.
3. Intangible cultural heritage as individual tacit knowledge
The connection between ICH and tacit knowledge becomes compelling, particularly after correcting an overly collective interpretation of the Convention.
3.1 Convergence of characteristics
| Characteristic | Tacit knowledge (Polanyi) | Individual ICH (UNESCO 2003) |
|---|---|---|
| Nature | Embodied, experiential | Embodied, lived |
| Transmission | Through practice, imitation, apprenticeship | Through oral transmission, observation, participation |
| Formalization | Resists exhaustive codification | Resists complete documentation |
| Identity | Constitutes the person's practical identity | Constitutes the person's cultural identity |
Individual ICH thus appears as the culturally situated dimension of tacit knowledge – that part of our embodied heritage that connects us to a history, a family, a community, without losing its anchor in singular experience.
3.2 The three circles of heritage
The Convention, correctly interpreted, delineates a complete cartography of human heritage experience:
- The individual circle: What is proper to a person, their intimate history, their singular inheritances (a family lullaby, a transmitted gesture, a recipe modified over generations).
- The collective circle: What is shared by an identifiable group (a traditional dance, a community ritual, a local artisanal know-how).
- The common circle: What belongs to humanity as a whole (universal techniques, foundational narratives, shared values).
The major innovation of the Convention is to place the individual at the heart of this system. It is not the common that descends upon the individual, nor the collective that absorbs them. It is the sovereign individual who, through their recognition, can irrigate the collective and feed the common.
4. AI facing intangible cultural heritage
4.1 What AI can capture
AI excels at processing explicit knowledge about ICH:
- Documentation of practices (texts, images, videos)
- Structural analysis of musical or choreographic patterns
- Transcription and translation of oral narratives
- Databases of artisanal techniques
It can massively ingest these materials and restitute them in the form of open knowledge – accessible, consultable, shareable.
4.2 What escapes AI
The tacit, living, embodied part of ICH fundamentally escapes it:
- The lived experience of a ritual
- The intuitive adjustment between participants
- The sensation of the right gesture
- The profound meaning for the individual or community
- The intimate history that binds a person to a practice
This is precisely the limit of AI when faced with tacit knowledge in general, and with individual ICH in particular.
4.3 The extractive model of AI as a threat to sovereignty
Contemporary AI operates according to an extractive and aggregative model:
- Massive data aspiration (texts, images, sounds, traces)
- Decomposition into elementary units (tokens, statistical patterns)
- Recomposition to generate new content
In doing so, it dissolves individual origin. A gesture learned from a grandparent, once absorbed into a model, becomes a statistical correlation among billions. The individual's sovereignty over "what is their heritage" is ignored because it is not technically legible to the machine.
The passage from free knowledge (modifiable, whose origin can be traced) to merely open knowledge (accessible but whose provenance is drowned) is here crucial. AI tends to transform the "free" (with its four freedoms) into the "open" (accessible but whose genealogy is lost).
5. A proposition: individual sovereignty as a foundation for an ethics of AI
The UNESCO Convention of 17 October 2003, a law adopted by 180 States, offers a conceptual and legal framework for thinking differently about the relationship between humans and AI.
5.1 The principle of individual sovereignty
If the individual is sovereign in recognizing what falls within their intangible cultural heritage, then:
- They are the legitimate depositary of this heritage, without necessary external validation.
- They have authority to control its use by third-party systems, including AI.
- They can choose to share it (to contribute it to the collective or the common) according to terms they define.
This sovereignty is primary: it does not derive from recognition by the State, the community, or the expert. It is inherent to the person.
5.2 Implications for the design of AI systems
This principle should guide the design of AI systems respectful of cultural rights, capable of:
- Recognizing provenance: Knowing that an element of knowledge comes from a sovereign human source, with its history and significance.
- Respecting integrity: Not denaturing a heritage element by treating it as mere interchangeable data.
- Enabling consent: Offering the individual the possibility to say "this is my heritage, I choose whether or not to share it, and under what conditions."
- Ensuring traceability: Even in recomposition, allowing the individual contribution to the result to be traced.
- Guaranteeing reversibility: Allowing the individual to withdraw their heritage from an AI system if they wish.
5.3 A hierarchy of requirements according to heritage circles
| Type of ICH | Relation to AI | Minimum ethical requirement |
|---|---|---|
| Common | May be harvested without individual consent, as by definition shared by humanity. | Ensure that the common remains common (no exclusive private appropriation). |
| Collective | Requires the consent of the community. | Define procedures for representative collective consent. |
| Individual | Requires the explicit consent of the sovereign individual. | Prior, informed, and revocable individual consent. |
6. Conclusion: towards an AI at the service of heritage sovereignty
The UNESCO Convention of 17 October 2003 is not an exotic text reserved for heritage specialists. It is, in many respects, a global law for the recognition of the human in their cultural singularity.
Its Article 2, correctly interpreted, offers a model for thinking about the place of the individual in relation to AI systems. It reminds us that:
- The human is not primarily a "data producer" but a bearer of heritage.
- Knowledge is not merely an "informational good" but a part of identity.
- Sovereignty is not only a matter of the State but a prerogative of the person.
Faced with AI systems designed to extract, aggregate, and recompose without regard for sources, this international legal framework invites us to reverse the perspective. The task would be to design AI systems at the service of individual and collective sovereignty – capable of recognizing, respecting, and valuing the part of humanity that each individual carries within them.
The 180 States that have ratified this Convention have, perhaps without knowing it, laid the foundations for an ethics of artificial intelligence based on cultural rights. It now falls to researchers, developers, and legislators to draw all the consequences from this framework to build a truly human AI.
Author: IRHM
Date: 19 February 2026
Version: 1.0
This paper was developed from an in-depth exchange on the relationships between open science, artificial intelligence, and intangible cultural heritage, in light of the UNESCO Convention of 17 October 2003.