Reliable AI, Risk-Sharing, and Equal Respect
This research line develops the argument that increasingly reliable predictive AI can threaten risk-sharing cooperation. When algorithms make individual risk profiles more legible, social institutions may shift from pooled uncertainty toward individualized risk pricing, weakening the moral space for solidarity and equal respect.
Status
This project develops from the presentation The Paradox of Reliable AI: How Optimized Algorithms Erode Risk-Sharing Cooperation, delivered at Existential Threats and Other Disasters: How Should We Address Them, Hotel & Resort Mediteran, Budva, Montenegro, 30-31 May 2024.
The current public page is an orientation to the developing article project, not a public preprint or final manuscript.
Core Questions
- Can more reliable prediction make risk-sharing institutions morally and politically harder to sustain?
- When does actuarial differentiation undermine equal respect among citizens?
- How should liberal and egalitarian theory respond when uncertainty is replaced by algorithmic profiling?
- Can information coarsening or use-based prohibitions preserve the moral space for brute-luck pooling?
- What should algorithmic governance protect besides accuracy, efficiency, and market transparency?
Orientation
- The project contrasts meritocratic cooperation with risk-sharing cooperation.
- It treats brute-luck pooling as an institutional expression of equal respect, not merely as actuarial convenience.
- It explores "ignorance-effect" regulation: preserving some effects of not knowing in domains where full risk differentiation would corrode solidarity.
Public Version Note
This page records the public-facing research orientation for a developing article line. If a public manuscript, accepted article, or English access edition is later posted, it should receive its own citation metadata and publication page.