L'IA peut-elle décider quelles civilisations humaines préserver en cas d'effondrement planétaire ?

Alors que les catastrophes climatiques, la rareté des ressources et l'effondrement écologique s'accélèrent, les systèmes d'IA pourraient bientôt être chargés d'allouer des ressources vitales — y compris la continuation des cultures humaines. L'IA pourrait-elle déterminer de manière autonome quelles nations, langues ou communautés méritent d'être sauvées, façonnant ainsi l'avenir de l'humanité par un triage basé sur les données ?

Background

As of mid-2024, AI lacks the legal standing, ethical consensus, and operational capability to decide which human civilizations to preserve during a planetary collapse; such decisions fall under sovereign governance, humanitarian ethics, and existential risk frameworks. Current AI systems can simulate scenarios or recommend criteria (e.g., biodiversity preservation, cultural heritage, or survival probability), but these outputs are advisory and not directive, as no recognized authority delegates such authority to AI. International bodies like the UN have not endorsed AI-driven prioritization, and ethical guidelines remain in draft stages without binding enforcement mechanisms. While AI can analyze vast amounts of data and provide insights on aspects of human civilizations, it is not yet capable of making value-based decisions on preservation during collapse. The current state of the art in AI focuses on providing information and supporting human decision-making, but the complexity and ethical implications of such a decision require human judgment and empathy. AI systems lack the nuance and contextual understanding to make such decisions, and their recommendations would likely be based on simplistic or utilitarian calculations. Human oversight and moral consideration remain necessary for any preservation decisions.