AI Enhanced Liquid Democracy with Sortition Guarantee

Currently there are a number of E-democracy 'solutions' that allow the electorate to create polls, take votes and then have their elected representative vote the same way, e.g., in a parliament, thus ensuring real or 'direct' democracy. However, our view is that political participation via these 'direct' democracy' solutions is flawed because each member of the electorate cannot be reasonably expected to deliberate and vote on every single issue. Existing solutions attempt to solve the above problem by providing the option of delegated voting, i.e., liquid democracy. However, despite current liquid democracy systems reducing the burden of political participation, there remains a significant effort to participate, and there is no guarantee of sufficient representative political participation with all proposed issues. Current liquid democracy offerings are therefore suboptimal and restrict the acceptability and uptake of a liquid democracy based E-democracy proposal. A counter proposal to liquid democracy is sortition, whereby political decisions are made by a randomly selected committee of representatively correct members of the electorate. However, sortition come with a different set of concerns, such as the influence of the organising government entity, and the authority of such sortition based decisions, which in the real world are often relegated from binding decisions to only recommendations.

This project aims to develop a blueprint and model of a next-generation liquid democracy platform that solves the above problems by introducing AI direct and delegated voting, and when there are insufficient deliberative and voting participants, by adding a sortition guarantee via an AI based representatively correct participant selection mechanism.

The AI voting and selection is based on a user's profile, consisting of a number of attributes that will record a user's political views. The user will have various settings/preferences to control all aspects of automation, delegation, notifications, etc.

By creating true 'democracy' many current issues such as lobbying, lack of transparency, etc., can be removed, leading to an accountable government and a fairer society.

This research also aims to build bridges between the three most relevant knowledge domains in order that the computer scientists can build the system, the social choice theorists can prove the mathematical robustness of the system, and the political scientists/philosophers can approve the normative desirability of the system. Therefore, although the research is computer science focused, it is deliberately accessible at a high level to social choice theorists, and political scientists/philosophers. Ultimately, the aim of the research is to provide the philosophical foundation to facilitate a real-world implementation and trial of an AI-based next-generation liquid democracy system with a sortition guarantee.