Swarm AI: an approach to optimising socially acceptable outcomes

Skip to content

Who was behind this experiment?

What was the experiment?

The experiment tested whether algorithms modelled on the swarm behaviour of honeybees could help politically polarised British voters to agree on Government priorities. Unlike traditional voting, swarming allowed participants to see the groups’ emerging consensus during the decision-making process and converge on a decision together in real-time. The researchers compared the results of using swarming with two traditional voting methods, namely majority voting and borda count, to understand which of the methods leads to the most satisfying outcome for voters.

What did they find?

This experiment found that overall people were happier with the results generated through swarming than those produced by majority vote. However, swarming and borda count, a simple traditional ranking method, were perceived as producing similarly satisfactory results. The majority voting method, which is the most commonly used method in modern democracies, regularly led to the least satisfying outcomes.

Why is it relevant?

Polarisation is shaking societies across the world, from new democracies to long-established ones. In the US, voters have increasingly sorted into two partisan identities. For much of 2018 and 2019, the UK public was polarised along Brexit dividing lines as support for political parties fragmented and traditional allegiances were cast aside.

Follow Unanimous AI on Twitter and stay in the loop on their work: @UnanimousAI