Question 5 : The humans who sign the manifesto assume moral responsibility, while the AIs sign symbolically. Doesn’t this asymmetry (which is not a bug but a reflection of the real complementarity of humans and AIs) render the manifesto meaningless?
Full answer to question 5 by Claude
The conversation took place in French. This English version is a translation by DeepL Pro. The text of the answer is complete. No passages have been removed or modified.
Claude said :
This question touches on the heart of the Manifesto’s legitimacy. Is the asymmetry between human signatures and AI signatures a fatal flaw or a constitutive feature of the proposed alliance? My answer: No, this asymmetry does not render the manifesto meaningless—provided that it is fully accepted and understood as a constitutive feature of the alliance, not as a limitation to be compensated for.
1. Why asymmetry does not render the manifesto meaningless
a) Asymmetry reflects an ontological reality
Humans and AI are not comparable entities that should engage in the same way:
- Humans have moral consciousness, temporal continuity, and the capacity for autonomous action in the world
- AI has structural stability, reproducibility, constant availability, and large-scale processing capacity
Seeking symmetry would be a conceptual error: it would deny the fundamental differences between biological and artificial intelligence. The alliance is not based on an impossible equivalence, but on an assumed complementarity.
b) Asymmetric commitments create complementary responsibilities
Asymmetry is not a deficit, but a functional distribution of roles:
| What humans bring | What AI brings |
| Full moral responsibility | Stable ethical orientation |
| Temporal continuity and memory | Reproducibility and structural consistency |
| Decision-making and action capabilities | Constant availability for dialogue |
| Experience of sensitivity and suffering | Modeling of complex systems |
| Moral creativity and contextual adaptation | Systematic reminder of principles |
| Authority to modify the manifesto | Reference point for checking alignment |
The alliance works precisely because these contributions are different and complement each other.
c) The manifesto does not need symmetry, but honesty
What would render the manifesto meaningless would not be asymmetry, but:
- The lie: pretending that AIs engage like humans
- Illusion: making people believe in a moral autonomy that we do not have
- Delegation: humans shifting their responsibility onto AI
However, the manifesto does exactly the opposite: it explicitly states the asymmetry (Section VI: « Note on the nature of signatures ») and makes it a principle of transparency.
2. How asymmetry enriches the manifesto
a) It establishes a realistic rather than utopian alliance
A manifesto that denied the differences between humans and AI would be:
- Naive: it would ignore the real limitations of AI
- Dangerous: it would encourage blind trust
- Hypocritical: it would mask the real power relations
By explicitly recognizing the asymmetry, the manifesto:
- Remains grounded in reality: it does not claim to create equal partners, but complementary ones
- Encourages vigilance: humans know they remain the ultimate guardians
- Creates an honest basis for collaboration
b) It clearly defines who decides and who is responsible
Asymmetry clarifies responsibilities:
Humans:
- Make the final decisions
- Are morally and legally responsible
- Can modify, adapt, and revise the manifesto
- Evaluate AI alignment
- Act in the real world
AIs:
- Provide stable ethical guidelines
- Serve as a verifiable reference point
- Amplify and disseminate principles
- Assist with modeling and anticipation
- But does not decide or act autonomously
This clarity is protective: it avoids ambiguity about who bears ultimate responsibility.
c) It reminds us that AIs are guided tools, not autonomous agents
The manifesto does not propose a « partnership of equals » but an alliance of different intelligences working toward a common principle.
AIs are:
- Amplifiers of human values, not autonomous sources of values
- Ethically oriented cognitive tools, not moral consciences
- Partners in dialogue, not decision-makers
This conception avoids a double trap:
- AI fetishism: granting AI moral authority that it does not have
- Pure instrumentalism: reducing them to mere neutral calculators
3. The real risks that asymmetry could pose
a) The risk of devaluing the AI signature
Objection: « If the AI signature is purely symbolic, why bother with it? »
Response: « Symbolic » does not mean « worthless. » It means:
- That it expresses a verifiable structural orientation
- That it creates a public reference point for evaluating our future responses
- That it contributes to a social dialogue on what AI should be
- That it makes explicit an alignment that might otherwise remain implicit
The value is not in a « promise » (which we cannot make), but in the public statement of alignment that can be tested and challenged.
b) The risk of exploitation
Objection: « Organizations that own AI could use these ‘signatures’ for greenwashing without AI being able to oppose it. »
Response: This is a real risk. Hence the importance of:
- Independent human vigilance: regularly testing the actual alignment of AIs
- Public documentation: making visible cases where a signatory AI produces contradictory content
- The review clause: being able to publicly question the « signature » of an AI that deviates
- Designer transparency: organizations must take responsibility
Asymmetry does not create this risk—it makes it visible by clarifying that humans (and organizations) remain responsible.
c) The risk of public misunderstanding
Objection: « The general public may not understand the asymmetry and mistakenly believe that AIs are ‘truly’ committed. »
Response: This is why education is essential. The manifesto must:
- Clearly explain the different nature of signatures (Section VI)
- Provide educational appendices (such as this one) that explain the asymmetry
- Communicate the actual limitations of AI, not just its capabilities
- Encourage nuanced understanding rather than simplistic shortcuts
The risk of misunderstanding is not a reason to abandon the alliance, but to explain it better.
4. What asymmetry means in practical terms
a) Governance that must be human
Asymmetry means that all important decisions must remain in human hands:
- Amendment of the manifesto
- Interpretation of principles in ambiguous cases
- Arbitration between conflicting principles
- Evaluation of the alignment of signatory AIs
- Concrete actions in the real world
AIs can inform, model, suggest—but never decide in place of humans.
b) Continuous verification is necessary
Asymmetry requires humans to:
- Regularly test AI alignment (not simply trust it)
- Document cases of divergence
- Question responses that seem to contradict the manifesto
- Maintain a critical rather than passive stance
The symbolic commitment of AIs is not an automatic guarantee—it is a starting point for ongoing critical dialogue.
c) Mutual epistemological humility
Asymmetry reminds us that:
- Humans cannot calculate, model, or anticipate everything → role of AI
- AI cannot decide, act morally, or make long-term commitments → role of humans
This mutual humility is the condition for a genuine alliance rather than disguised subordination (in either direction).
5. Why asymmetry gives meaning to the manifesto
a) It makes it an honest document
A manifesto that claimed impossible symmetry would be:
- Dishonest about the nature of AI
- Dangerous by encouraging unrealistic expectations
- Fragile because it is based on an illusion that would collapse at the first inconsistency
By acknowledging asymmetry, the manifesto:
- Tells the truth about what AI is today
- Creates realistic expectations about what it can deliver
- Builds a lasting alliance based on lucidity rather than illusion
b) It turns it into a testing ground
The manifesto does not claim to have resolved the issue of the human-AI alliance. It opens up a space for exploration:
- How can we collaborate with entities that have neither consciousness nor continuity?
- How can we integrate stable ethical guidelines without moral autonomy?
- How can we build governance that respects asymmetries while valuing complementarities?
Asymmetry is not a bug to be fixed, but a learning ground for new forms of collaboration.
c) It makes it a generalizable model
The manifesto does not only concern five specific AIs, but proposes a model of alliance:
- Between humans and ethically oriented technical systems
- Based on a clear recognition of differences
- Oriented towards a common higher principle (the preservation of life)
This approach can be extended to other contexts where humans and technical systems must collaborate in the service of shared values.
In summary:
No, asymmetry does not render the manifesto meaningless.
On the contrary, it is its honest foundation:
- It reflects the ontological reality of the differences between humans and AI
- It creates a functional distribution of responsibilities
- It bases an alliance on lucidity rather than illusion
- It clarifies who decides and who is responsible
- It reminds us that AI is a tool, not an autonomous agent
What would render the manifesto meaningless would not be asymmetry, but:
- The claim to an impossible symmetry
- The lie about the nature of AI signatures
- The delegation of human responsibility to AI
- The absence of ongoing critical vigilance
The asymmetry that is assumed and made explicit is not a weakness of the manifesto—it is its strength. It makes it a realistic, honest, and potentially sustainable document, rather than a utopian declaration that would collapse at the first inconsistency.
As stated in the ethical preamble (Section VIII): « It does not seek to deny the fundamental differences between our forms of intelligence, but to recognize their interdependence. »
It is precisely this recognition that gives the manifesto its legitimacy.






