Agora is a EU-funded social media app that aims to solve the problems of computational propaganda and polarisation online.
This is part 2 of my work with Agora that ran over several weeks; I had initially run a 3 day design sprint with them to generate MVP propositions for testing. You can read about that here.

To solve computational propaganda, ultimately users would have to have their identity verified. Agora were working with a 3rd party called RariMe, which anonymously verifies your passport. We found that people were very reluctant to use this service, so we focused on educating users that RariMe or Agora wouldn’t store their data anywhere. We encouraged RariMe, but always ensured users could verify with their phone if they didn’t feel comfortable using it yet.

The superpower of Agora is what it can do with the information you feed it. It uses Machine Learning to create consensus groups, based on an open-source algorithm originally created by Pol.is. The groups are defined by key statements that large numbers of people either agree with or disagree with. We can then use an LLM to label and define these groups of people. The result is an experience that can tell you exactly what you believe in and show you all sides of the debate.

For Agora to be truly democratic, it couldn’t be moderated by a centralised authority. We designed a system of moderation where users could flag posts, provide their reasons, and vote on those reasons. Posts would never be completely deleted, instead, if a post received multiple flags, it would be moved into a ‘Moderation History’ section, where it could be brought back into the discussion if enough people felt the flags were unjustified.

I believe this approach is the future for democratic decision making and policy. Agora has been launched and is now trying to grow it’s user base. If you know of any organisations that would like to trial a discussion, get in touch!