17 February 2015

A rational reconstruction of democratic thinking

How can we use technology to make democracy better?
  • Can we make quicker and better decisions?
    • The choices are those that we would come to make if we all thought long and deeply and did a lot of research; but they'd happen quicker than they do today.
  • Can we make taking part in democracy a good experience?
    • It should be educative, empowering, and fun for participants.


For any topic, there is a relatively small number of people who think deeply about the topic, discuss it, and vote in a way that represents what would happen if the whole population engaged in the same exercise of deep thinking and discussion.

Concept: democracy by entrusting thinkers

People can entrust others to think and act on their behalf on specific topics. If you entrust someone on a topic, they are empowered to tag issues with the topic, and if you do not vote on the tagged issue, they can vote on your behalf. If you entrust several people on a topic, your vote is split equally between them.


  • There are issues that people discuss. Things like immigration policy, when the bins should be picked up, how to spend local council money, what language to write some bit of open-source software in, etc.
  • Issues can be tagged with topics of expertise needed to make a decision on the issue.
  • People can vote on issues.
  • People can delegate to each other on topics.

Important Questions

  • How are topics selected? Should people propose topics they can be delegated to?
  • Tagging lead time? Should there be a time delay after someone you entrust tags a topic before they can cast a vote on your behalf (giving you a chance to see what they tag first)? It could be user and issue configurable.
  • Timing issues? Should there be a time delay for a delegated vote so that you can see how someone is voting on your behalf and change your mind? It could be user and issue configurable.
  • Entrustment timing? Should we allow people to change trust at any time?
  • Basis for entrusting others? Should entrusting to someone enable you to see what they do on your behalf? Can we make a good platform for expressing and summarising people's arguments?
  • Anonymity? Should anonymity of voting be preserved unless they the voter has specified that they can be entrusted to?
  • Verifiability? Should it be possible to verify your vote was counted correctly?
  • What happens if I change who I trust after a vote? If a vote has happened, and someone has acted on your behalf for that vote, you cannot change it after the votes have been cast.
  • Instead of equal split of trust between people, it could be weighted, or stack-ranked.
  • One can do clever things with tags: e.g. I trust you on your tag X (e.g. software engineering), with my tag Y (computer science).
  • How will the system try to be gamed? How to defend against it? Analysis of the effectiveness of gaming is needed.
  • How can views between tags - networking topics - be done consistently and in a way that helps people navigate the topics? What is the impact of incorrect tagging?
  • How can the process be analysed by social scientists live for fairness, social influence etc (eg see http://news.sciencemag.org/2012/06/who-controls-social-networks)
    • Would such a system change the way people act in a bad way and in particular, would it be worse than political corruption today?
    • Observation: the proposed system is much more transparent, and allows anyone to be involved, so we would expect less corruption.
    • What impact will reputation system have on participants? Will they be incentivised to have people entrust them? Will this be a good or a bad thing?


  • New political parties that will act according to fluid trust (e.g. something like http://democracyos.org).
  • Participatory budgeting allocation (do we need to build modern decentralized tools to monitor and build reports?).
  • Businesses and other organizations (co-ops, ngos, companies, etc) internal organization structure.
  • Consultative processes in existing democracies.