Skip to content
Swarm AI diagram

A diagram explaining Swarm AI

Who is behind this experiment?

What is the experiment?

The experiment aims to test whether algorithms modelled on ‘swarm’ behaviour in bees and fish can enable groups with conflicting political views to find collectively acceptable solutions. It will compare crowdsourced decision-making using Swarm AI technology against traditional forms of decision-making including majority voting, ranked voting, and pairwise voting. The experiment will test whether Swarm AI can optimise ‘overall group satisfaction’ with the decisions made. Participants will be tasked with prioritising government objectives on issues that are often divisive.

Why is it relevant?

In an increasingly polarised world, we need to find new ways of helping people with opposing views to find ways to agree. With the scale and complexity of challenges - from climate change to cyber threats - we need new tools to facilitate collaboration and cooperation. Recent experience has shown that traditional voting mechanisms, such as referenda, can sow greater polarisation of views. This experiment will test whether an AI-enabled approach can reduce dissatisfaction with group decisions better than our usual voting methods.

How might the findings help people better design collective intelligence?

These experiment findings will be helpful for anyone looking to design better ways of enabling groups of citizens to make decisions on highly-charged or polarising topics in a way that maximises the satisfaction of participants with the end result. It may have relevance for government and local authorities and civil society institutions who traditionally turn to opinion polling or surveying to understand citizen preferences and decide priorities.

Follow Unanimous AI on Twitter and stay in the loop on the experiment and swarms more broadly: @UnanimousAI