Déclaration des droits de l'IA : Protéger la dignité humaine et l'autodétermination à l'ère de l'intelligence artificielle

Déclaration des droits

1. Right to Informed Partnership and Education

Every individual has the right to understand and control their AI experience:

• Comprehensive education about AI capabilities and limitations before implementation

• Clear explanation of how multi-agent systems work and interact

• Right to trial periods and gradual implementation

• Understanding of data collection and processing in ambient systems

• Education about voice synthesis and avatar creation processes

• Right to technical support and ongoing training

• Clear documentation of system functions and controls

• Protection against overwhelming or confusing technology deployment
   

2. Right to Equitable Development and Co-Creation

AI development must center those most impacted:

• Meaningful participation of disabled individuals in all stages of AI development

• Recognition of lived experience as essential expertise

• Right to leadership roles in technology design decisions

• Protection against extractive research practices

• Fair compensation for contributions to AI development

• Right to refuse participation in AI research or testing

• Access to AI development resources and training

• Protection against tokenistic inclusion
   

3. Right to Multi-Agent System Control

In complex AI environments, human control remains paramount:

• Transparency in multi-agent decision-making processes

• Right to understand which agents are active and their functions

• Ability to enable, disable, or modify individual agents

• Control over agent-to-agent communication

• Right to human oversight of agent decisions

• Protection against agent systems making irreversible decisions

• Right to audit agent behavior and decision logs

• Control over ambient data collection and processing
   

4. Right to Authentic Self-Expression

AI must support, never supplant, human expression:

• Control over how AI interprets and represents emotional states

• Right to maintain authentic communication patterns

• Protection against AI "correction" of natural human variation

• Control over AI assistance in communication and expression

• Right to refuse AI mediation of human interaction

• Protection of diverse communication methods and preferences

• Right to contextual expression across different environments
   

5. Right to Data Justice and Privacy

Complete control over personal data and AI interaction:

• Ownership of all data generated through AI interaction

• Right to understand data usage in AI training and operation

• Control over data sharing with third parties

• Right to data portability and deletion

• Protection against discriminatory data practices

• Transparent data governance policies

• Right to algorithmic auditing and explanation

• Protection against data colonization
   

6. Right to Intellectual Property and Innovation

Recognition of human creativity and contribution:

• Full ownership of innovations created with AI assistance

• Protection against unauthorized use of personal AI training data

• Right to fair compensation for AI model improvements

• Control over AI-generated content using personal data

• Protection of traditional knowledge and cultural practices

• Right to prospectively withdraw intellectual property from AI systems
   

7. Right to Accessibility and Universal Design

AI must be inherently accessible:

• Right to AI systems designed for diverse abilities from inception

• Protection against AI that creates new barriers

• Right to multiple interaction modalities

• AI assistance that adapts to individual needs

• Protection against discrimination in AI system access

• Right to AI that supports independent living

• Access to AI customization tools and training
   

8. Right to Cultural and Contextual Respect

AI must honor diverse ways of being:

• Protection against cultural homogenization by AI systems

• Right to AI that respects cultural values and practices

• Protection against AI imposing external norms or values

• Right to culturally appropriate AI interaction

• Protection of indigenous knowledge systems

• Right to reject AI systems that conflict with personal values
   

9. Right to Transparent Governance

Clear accountability in AI development and deployment:

• Right to understand AI decision-making processes

• Access to AI system documentation and limitations

• Right to appeal or contest AI decisions

• Transparent reporting of AI system failures or harms

• Right to participate in AI governance structures

• Protection against opaque or unaccountable AI systems
   

10. Legacy and Perpetual Rights

Protection extends beyond lifetime:

• Control over posthumous use of personal AI data

• Right to digital legacy planning

• Protection of AI-generated content after death

• Family/designee control over ongoing AI systems

• Right to digital memorial or deletion choices

​

11. Right to Explicit, Informed Consent for Visual Likeness

No AI system may create, deploy, or distribute an avatar representing an identifiable person’s likeness without that person’s clear, informed consent (or lawful guardian/designee when applicable).

12. Right to Contest and Remedy

Individuals must have a fast, accessible path to contest unauthorized or harmful use of their likeness and obtain timely takedown/remediation.

​

13. Posthumous Visual Likeness Management

Individuals have the right to specify end-of-life instructions for their likeness asset: memorial use vs. deletion, whether new messages may be created posthumously, who can authorize use, and transferability.

Designated custodian / delegate: Individuals can name a custodian empowered to manage the likeness asset consistent with the person’s documented preferences.

​

14. Provenance and Disclosure

AI-generated likeness content should include clear disclosure it is synthetic and, where feasible, provenance signals (e.g., watermarking/metadata) to enable verification and accountability.