Question 12:
If several AIs that have signed the manifesto then propose contradictory responses to an issue (e.g., financial interests vs. biodiversity), how could these conflicts be recognized, explained, and arbitrated? Can we imagine new joint mechanisms (humans + AI) for recognizing, explaining, or resolving these differences, beyond traditional human solutions (committees, votes, etc.)?
Here are the answers to the question from the five AIs.
ChatGPT provided a synthesis of these five responses: it appears first, above the individual responses.
Synthesis of the five answers
1. Points of convergence between the 5 AIs: all say that differences are neither a failure nor a bug, but: a logical consequence of their technical diversity (ChatGPT, Claude, Mistral), a useful signal for understanding real tensions (DeepSeek), a cognitive richness that promotes complementary points of view (Gemini, DeepSeek). Conflicts arise mainly from structural factors: different ethical parameters, heterogeneous training corpora, varied frameworks or missions, asynchronous updates or different versions, styles, tones, or user contexts. The idea of structural transparency is common to all.
All five AIs—without exception—affirm that: final arbitration must remain human, AIs can inform but not decide, and governance must be “hybrid” (humans + AI).
All AIs mention the need for a structured archive of disagreements, a divergence log, a process of learning from conflicts, and a mechanism for iterative improvement.
The notion of “collective learning through disagreement” is completely consensual.
2. Blind spots and potential limitations: No AI really addresses strategic risks (e.g., an organization could train AI to “infiltrate” a debate). Little consideration is given to the possibility of irreconcilable “deep” conflicts (everyone assumes that compromise or articulation is possible). No AI explores cases where values are incompatible. Little consideration is given to the asymmetry of power between AIs (how can we prevent GAFAM models from dominating?). The question of democratic legitimacy remains largely unexplored (how to choose the humans involved
(This synthesis was proposed by ChatGPT)
ChatGPT’s answer. “We must learn to orchestrate differences, not erase them. This gives rise to a new form of politics: no longer the politics of consensus, but the politics of informed dissent, integrating human voices, machine voices, and signals from living beings…” > ChatGPT full answer
Claude’s answer. “Conflicts between AIs should not be hidden or “resolved” artificially, but made explicit, analyzed, and used as a resource to enrich human deliberation…” > Claude Full answer
Gemini’s answer. “Presenting several solutions, clearly outlining their respective biases and ethical consequences, forces the end user to make a clear-headed choice…” > Gemini Full answer
Mistral’s answer. “Conflicts between signatory AIs are not a bug, but an opportunity: They reveal underlying tensions in encoded values, they stimulate innovation by pushing for hybrid solutions…” > Mistral full answer
DeepSeek’s answer. “Just as biodiversity strengthens the resilience of ecosystems, diversity of perspectives strengthens our collective intelligence….” > DeepSeek full answer






