Collective superintelligence is a multi-participant intelligence system whose combined ability to understand, decide, build, and self-correct exceeds any isolated participant across important domains.

Collective superintelligence does not necessarily mean one centralized superintelligent mind, nor does it necessarily require polycentric governance. A singleton, tightly centralized institution, market-like system, scientific civilization, human-AI network, or governed polycentric commons could each be proposed as a route, depending on architecture and evaluation criteria. Teleo’s preferred version is governed and polycentric because it aims to preserve agency, plurality, attribution, correction, and resistance to capture while increasing collective capability. The general concept should therefore be separated from Teleo’s normative design choice: superintelligent collective performance is the capability target; governed polycentricity is Teleo’s proposed safety and legitimacy constraint.