Methods to scale deliberative decision-making to billions of simultaneous people

In today’s connected information environment, communication with millions of individuals is the norm rather than the exception. However, communication is primarily text-based, asynchronous and non-deliberative. Real-time deliberation remains sporadic. Even on social media, where people can simultaneously signal their opinions to millions of individuals and engage in lengthy discussion threads, communication remains limited to broadcasting, and information aggregation is not guaranteed in a reasonable amount of time. Here, we present a novel approach to collective decision-making for open-ended problems, which scales real-time online deliberation to virtually thousands, millions, or even billions of individuals. The method, called PSi (People Supported intelligence), is an iterative method for group discussions that outputs a ranking of crowdsourced answers to open-ended questions. PSi is bounded by the number of iterations required to reach a collective decision and, thus, the time needed to reach it. It is guaranteed to output a decision in logarithmic time as a function of group size, opening the door to massive-scale deliberative decision-making. We present results from a discussion between Republicans and Democrats (N=208) on abortion rights. We present the PSi method within the broader field of collective intelligence and decision-making, its informational properties, and applications. PSi represents a scalable solution to deliberative decision-making that is theoretically resilient to herding, negative social influence, or intentional attacks (e.g., disinformation). We discuss its potential impact on deliberative democracy and deliberative participation in large groups, like organizations, cities, and societies.