A Confidence-Adjusted Consensus Mechanism for Scalable Deliberative Decision-Making

As groups grow in size, the complexity of collective decision-making increases exponentially, leading societies to concentrate authority in small hierarchical subgroups. While this enables coordination at scale, it systematically underrepresents the interests of broader populations and limits collective learning. Traditional voting mechanisms exacerbate these problems: binary choices, fixed option sets, and winner-take-all outcomes incentivize polarization rather than consensus-seeking. This paper introduces a deliberative framework designed to support meaningful participation at scale. The framework combines three innovations: (1) open and continuous proposal generation, allowing participants to introduce new alternatives throughout the process; (2) continuous preference expression on a scale from strong opposition to strong support; and (3) real-time aggregation using the Consensus Algorithm, which calculates scores as Mean − SEM (standard error of the mean), yielding a confidence-adjusted estimate of collective agreement that penalizes uncertainty and protects minority positions. Two proof-of-concept applications demonstrate the framework's practical viability. In the first, 53 participants converged on a single name from 26 proposals within approximately five minutes. In the second, 40 participants representing secular and religious perspectives jointly developed a social charter on religion-state relations over two sessions totaling 5 hours, achieving consensus above 60% on key provisions. These results suggest that appropriately designed deliberative technologies can enable rapid, inclusive decision-making on both simple and normatively contested issues—offering a pathway toward scalable deliberative democracy.