Sortition
Sortition is the practice of selecting participants in governance institutions by random draw rather than by election, volunteer, or appointment. In modern participatory contexts, the random selection is typically stratified to ensure demographic representativeness across age, gender, race, income, education, and geography. Fung identifies sortition as the essential response to the self-selection bias that degrades most open participation: any process requiring volunteers systematically underrepresents populations that lack prior political engagement, precisely the populations whose inclusion participatory governance most requires. The Irish citizens' assemblies, French climate convention, and Canadian citizens' assemblies on electoral reform demonstrated that sortition produces participant pools more representative of the general population than any self-selected process.
In the AI Story
The case for sortition rests on evidence that self-selection produces systematically unrepresentative participation. Open public meetings attract professional advocates, organized interests, and highly engaged citizens whose perspectives are unlikely to match those of the general population. Public comment periods privilege technically literate respondents comfortable with regulatory procedure. Town halls draw those with time, transportation, and political identity. Each mechanism selects for characteristics that correlate with existing advantage and against the silent middle whose considered judgments participatory governance is designed to access.
Sortition addresses the problem directly: randomly selected citizens cannot be pre-organized, cannot be dominated by professional advocacy, and cannot be captured by organized interests. The specific form — stratified random selection from demographic databases — has been refined through multiple implementations. The Irish citizens' assemblies used a three-stage process combining demographic stratification, geographic representation, and voluntary acceptance of selection to produce panels matching the Irish population on major demographic dimensions.
Application to AI governance is straightforward. AI Impact Assemblies, National AI Deliberation Platforms, and other participatory mechanisms require selection processes that capture populations affected by AI without requiring those populations to opt in. Sortition's role is to convert demographic databases into governance-ready participant pools — a technical and administrative challenge but not a theoretical or political one.
The legitimacy of sortition faces cultural rather than technical obstacles. Elections are the familiar mode of democratic selection; random selection feels unfamiliar despite its Athenian pedigree. Fung's response is empirical: the evidence from implemented minipublics shows that sortition produces outcomes with strong democratic legitimacy when the participatory process meets the other conditions of empowered participatory governance. The unfamiliarity is cultural, not structural, and diminishes as implementation accumulates.
Origin
Sortition has ancient roots — Athenian democracy used random selection for most political offices — but largely disappeared from democratic practice with the rise of representative government. Its revival in the late twentieth century came through the work of Bernard Manin, John Burnheim, and others who argued that representative democracy had drifted from its democratic origins toward a form of elected oligarchy.
Practical implementation accelerated in the 2000s through experiments in British Columbia, the Netherlands, and Ireland. Each demonstrated that sortition could produce governance bodies with legitimacy comparable to elected institutions for specific governance tasks, particularly constitutional and policy questions where electoral incentives produced distorted outcomes.
Key Ideas
Self-selection is not neutral. Any participatory process requiring volunteers systematically overrepresents the already-engaged and underrepresents the silent middle.
Stratified random selection produces representativeness. Properly designed sortition generates participant pools matching the general population on demographic dimensions that correlate with governance-relevant perspectives.
Capture-resistance is the structural advantage. Randomly selected participants cannot be pre-organized by interest groups, providing resilience against the lobbying dynamics that distort other participatory mechanisms.
Legitimacy emerges from implementation. The cultural unfamiliarity of sortition diminishes as successful implementations accumulate and citizens observe its outcomes.
Appears in the Orange Pill Cycle
Further reading
- Bernard Manin, The Principles of Representative Government (Cambridge University Press, 1997)
- John Gastil and Erik Olin Wright, Legislature by Lot (Verso, 2019)
- Hélène Landemore, Open Democracy (Princeton University Press, 2020)