Designing Online Democracy - Some thoughts

I want to start a wide ranging discussion about Online Democracy solutions and what is really required for them to be widely successful.

Some background on me as it relates to this issue:

  • It probably helps to understand that I’ve been in software engineering for around 35 years, much of that in software architecture (large scale-able distributed designs) or research roles.
  • I wrote down a fairly condensed set of principles and ideas around the workings of a participatory democracy system a while back: I’m not hung up on the details of that too much, but reading it will give you some concept of where I’m starting from.
  • I’ve been a member of Pirate Party Australia (PPAU) for over 5 years, 4 years working on Policy Development, 2 of those as elected Policy Development Officer, and this years I am an elected member of the National Council. I think that PPAU does favour participatory democracy as a core value, even if there is no broad consensus on what eventual form that should take. The platform for my election to NC was largely about this.

My largest core issue with all attempts I have seen to date (including Flux, MiVote and Online Direct Democracy (prev k.a. Senator Online), and Pirate Party Germany’s Liquid Democracy solution), is about their structure. They all seem to me to embody similar structural failings. They tend to focus primarily on the issue of giving everyone a vote.

Voting is important, but in any collective decision making system, voting is just the formality that happens at the end when all the consensus building work is done.
It’s like when you make team decisions in a business. Decisions are not made at meetings. They get made before the meeting. The meeting is a formality, as is voting.
There are many reasons why this should be the case, not least amongst them, are predictability, trust and evidence of commitment to action.

Presenting voting as a inflection point in time where a binding decision is suddenly made, makes it a target for corruption.

  • Groups with agendas can swoop in at the last moment and overthrow all expectations without ever participating up to that point, instantly destroying predictability and trust. Such actions would be entirely valid as a purely vote exercising concern, and yet utterly destabilising as an institutional trust concern. If you want to see this in action in social media, just watch various sub-Reddit’s “brigading” each other.
  • There’s is also the well known effect whereby meetings are controlled by the people who write the agenda. In ‘mivote’ for example, the “impartial” people writing the policy briefings would be an amazingly powerful vector for manipulation. It doesn’t even have to be overt. It may even happen accidentally. It can just be in how the questions are posed. For a classic example of this, look at the way that John Howard scuttled the Republic referendum, just by the structure of the question he posed.

We have watched as the Pirate Party in Germany floundered around the use of their Liquid Democracy system. Here’s a description: It’s not really in use now. It made some stupid decisions. People lost faith. In their case, I think the problem was that the system tended towards popularity based voting. People mindlessly “jumped on bandwagons”, despite the relative simplicity of informing themselves.

Other examples of related failure modes include getup and These things flared up for a while as novelties, looking like people power in action, but in effect they are now viewed by the establishment as superficial. A getup petition has little political consequence, because voting in getup is too easy. It doesn’t carry conviction. A politician may see 100,000 votes on a petition from them. In the early days, the politicians didn’t know how to gauge that, because there was no precedence for understanding how much those voters cared about the issue. Now they know, and the answer is that it’s not a good indicator. People sending in their own emails is slightly better, if they’re not all form letters. People sending in actual hand written letters is far more significant, because such letters indicate that those people actually took time out of their lives to take action on the matter, put pen to paper and do the snail mail thing.

Click to vote is too easy. There has to be persistent visibility of people striving to comprehend the issues and their personal commitment to the result.
Commitment to act. Commitment to change governments if necessary. Without this, the simple act of voting means next to nothing.
This is the “evidence of commitment to action” bit.

Another important factor is the human tendency towards tribal grouping. People want to join groups of people like themselves.
Getting beyond the techno-nerds and people who think in abstract about the merits of open democratic process, and out to the general populace, will be damned near impossible unless they can join with groups they feel comfortable with. They don’t need to stay comfortable. We don’t want filter bubbles, but we do need them to join. This is a conundrum.

The obvious question at this point, is: What do I think a real solution would look like?
I’ll describe this by relating and contrasting it to some well known existing solutions.
I’m imaging a system that looks like a hybrid of Wikipedia, Reddit and Flux. The key question then being which aspects of each am I referring to?

  • The significant aspects of Wikipedia, are in the way that it produces a complex persistent, yet evolving, knowledge artefact, and the way it keeps history.

  • In Wikipedia’s case, there is only one answer and they strive mightily to take the neutral stance and to be as objective as can be.

  • By contrast, in this system,

    • Many concurrent “complex persistent, yet evolving, knowledge artefact”'s would need to coexist, representing the current consensus positions of groups of citizens.
    • They would be linked to alternative consensus positions by other groups of citizens, forming the basis of political debate within the system.
    • Like Wikipedia, there should also be history, so we have a record of how and why consensus changed over time.
  • The significant aspects of Reddit, are the fluid discussions maintain supported by up/down voting of content driving rampant discussion in many threads, and their identity groups in the form of sub-Reddit’s.

  • In Reddit’s case, the discussions are not grounded in any body of group consensus (except loosely in side docs or imposed by sub-Reddit mods)

  • By contrast, in this system,

    • Discussion threads would need to be grounded in refinement or extension of elements of current group consensus, or to contrast them against positions held by other consensus groups. There is an objective, to find agreement and document it.
    • The opportunity exists here to deliberately push back the filter bubble effect, by driving conversations about finding common ground and resolving misunderstandings between consensus groups.
  • The significant aspect of Flux, is in the way it embraces the distributed trust mechanisms enabled by block-chain technologies.

  • In Flux’s case, block-chain voting solutions are utilised to remove the need for any central trusted authority that may be corrupted.

  • By contrast, in this system,

    • The distribution of trust should be extended to incorporate all aspect of consensus building and the collective artefacts they produce.
    • The only way to alter consensus positions MUST be through convincing the majority of people that they should do so.
    • Pseudonymity is also important. One person = one vote, but without exposing them to IRL attack.

In this system,

  • A binding vote is really a globally verifiable snapshot of consensus at a point in time. No surprises.
  • Attractive identity (tribal) groups get to exist, and yet by design they are pushed out of their comfort zones and exposed to alternatives to help spread consensus.
  • Evidence and history will exist for anybody to see the extent of deliberation and consensus behind any externally proposed stance. Politicians would come to understand the demonstrated commitment this embodies.
  • Opportunities for corruption are eliminated by the widely distributed nature of the consensus process. Millions of micro-“attack surfaces” instead or a few large ones.

Excellent write up. Best I have read.

Trust and who is entitled to vote is the core issues imo.

Additional questions by Christian Tan. (About DODTDG in general):

Suggestions for future discussions:

  • What is the purpose or what are the intended outcomes of this group?

  • The scientific method has been brought up as a potential solution - what would such a democracy, built upon the scientific method, look like?

  • Regarding ‘finding solutions’ (i.e. discussing what an ‘ideal system’ would look like), three potential things to consider are:

    • the problem (what is it that ‘the solution’ aims to address),
    • the context (i.e. social and cultural context, which may differ in terms of having specifics relating to particular societies, such as Australia) and
    • the existing (what solutions are already out there i.e. certain forms of democracy - and their pros and cons)
  • The role that people (i.e. in the sense of human nature / the human condition) play in shaping a democracy (or any other system)


Before the scientific method gets applied there has to be an underlying morality. I don’t think there is a scientifically determined system of ethics. Questions like: ‘What do we believe the role of government to be?’, ‘What is the goal (if any) of society?’ ‘What outcomes do we desire?’ need to be discussed before we go looking for solutions.

Once we know what we are trying to do, then the scientific method comes into play.

Pirate morality in my mind is about equality of opportunity, autonomy to live your own life and the freedom to pursue happiness, as long as you aren’t impeding on the rights of others. From that starting point we can begin to address issues.

For example: Our Universal basic income through reverse taxation policy removes the punishing and onerous experience of dealing with Centrelink, makes survival on welfare easier, distributes it equitably by not punishing someone for working part-time, which is a flaw in the current system.

If we applied different morality, say for example we believed as a group that the state should only help those who contribute to it, we would want to abolish all welfare because the leeches on welfare deserve nothing.

1 Like

Yes indeed, the scientific method is the most effective meta tool we have for getting to the truth about things, but has virtually zero to say about which truths we should seek.

Any participatory democracy system design needs to have this understanding built in. Most of the consensus building processes will be about finding the commonalities in peoples values. From there, we can proceed to solutions and maybe utilize some good science along the way. When we get stuck in disagreement again, it will usually be about some details of our relative value systems that we need to go back to resolve.


Scientific method is useful to advise and to understand - but it is not and never has been a method of governance.
Science informed rather than science governed IMO is what would work.

This is inline with Frew’s reasoning above.

Ethics and Morality are important to political governance, but they are not everything. Although we can choose to make them the most important thing - but even then some ethics or morals will be held to higher account than others. Do we deport people for stealing a mars bar? I mean theft is a serious moral issue. Do we jail someone for using unlicenced music in youtube video they posted about their family holiday? Do we dismiss someone from office for arguably lying about their expenses claims for travel? - there is a slope that is a little slippery about morals and ethics.

For defining the “solution” - the approach identified above in my profession is called a solution architected approach. If you follow those steps the outcome will be narrow and whilst may be brilliant will struggle to gain and sustain traction in the broader ecology of governance and leadership.

Building the best moustrap in and of itself does not mean it will ever catch a mouse. It needs to be known, popular, cheaper, a nice colour, easy to buy, have positive ratings on Yelp or wherever and possibly PewdiePie doing a youtube about using it that is difficult to understand but all the teenagers go crazy about it.

You need an enterprise architecture, the solution approach will just lead to what has happened before.

A decomposition of each of the below related models will give the detail and understanding needed to build this.
Strategy -> Broader Business Ecology (Markets, Services, Partners, Competitors) -> Capabilities (People, Processes, Tech that inter-operate to create outcomes) -> Information -> Locations

The trick is not to get the model right. The trick is to build it, then use it, then it will progressively get better over time as it is challenged, improved and innovated.

The immediate rejection point is that most people will not see digital democracy as a business. It is. All ideas are. It is desperately fighting for survival against competitors.

1 Like

The system of democracy that is in place in Australia can trace its origins back through time (a biased but thoroughly entertaining history is Churchill’s " A History of the English Speaking Peoples"). I mention this for two reasons: 1. Working democratic processes are evolutionary and 2. People don’t like experimenting with processes that have real world consequences i.e. will I be able to go to the Hospital next week, or will there be a pay-wall. So it is unlikely that a form of computer based direct democracy is likely to gain much popularity any time soon. Its certainly unlikely to gain support from the Libs or the Labs for whom such proposals might be considered to be apostasy. However, there is every reason to experiment within the party. To demonstrate that policy and implementation of policy can be affected directly by the party membership and that this can work in an effective way, producing concensus and unity and smart decisions. i say this because I believe that whilst we should have a critique of voting systems, I’m dubious about the value of the Pirate Party building one or becoming bogged down in such a debate until we can demonstrate some influence of actually implementing it. As indicated above, the NVB would fail should the elected person just opt out as Madigan did. More important at the moment (IMHO) is to demonstrate how the Pirate Party is internally and directly democratic.

1 Like

We have no intention of trying to implement any online voting system for the Country until we have tested it for a long time. We aren’t even planning to use it to replace our annual Congress voting until we are happy that it works well and is better than the current system.

As @AndrewDowning points out in the OP, the problem with current voting systems is that they are just a means of carrying out a vote and don’t work to build consensus. Pirate Party Germany developed Liquid Feedback and implemented it as their internal voting system. Within six months issues became apparent with the system and led to a lot of infighting and criticism. Other Pirate Parties jumped on the bandwagon and we had more cases to study what worked and what didn’t.

Many people were happy to delegate their vote, but not pay any attention to where their vote went. This led to a situation in one country, where six people figured out that they had the majority of delegated votes between them, then used that control to run the Party as a gang of six. Active members who disagreed with the direction of the Party had no means to challenge the will of the vote holders and got demoralised and stopped helping and the Parties’ prospects declined.

This has made us quite wary of replacing formal voting at Congress with an online voting tool until it has been thoroughly tested. We would implement the system and use it to generate motions for Congress which would carry on like normal. Once we could see that it was working properly we can look at replacing our own decision making process with the online democracy we are planning to develop. If it works great for us, then we can look at using it on a larger scale.


Yeah delegating votes is a terrible idea and I think one of the major fault points of representative democracy. Nice to see such a brutally magnificent case of it imploding.

1 issue - 1 person - 1 vote… good enough to work for youtube videos good enough for democracy.
Main concerns with digital voting is the same as any online polling, the technical question of how do you make verifiable votes to stop fake/duplicate gaming of the system?

There was a fantastic presentation at about trust/security/verifiability in election software:

For your information there is a third meeting on 6:00 PM to 8:00 PM Wednesday, February 15, 2017 located in the MiVote office 5 Hoddle St, Collingwood, Melbourne. This meeting will have Max Kaye from Flux party (voting reform party) present.

1 Like

fyi … via MorpheusB on irc.

Here is a good read about how representative democracy is failing and why.

Society is too complex to have a President complex mathematics suggests

Some thoughts and questions:

  • The Nation State is probably as redundant as Presidents, not much we can do about that at the moment. (Probably worth keeping in the back of our minds though).

  • How do we empower experts in specific fields democratically? Giving government over to experts without democratic control would be as dystopian as electing a Russian-backed orange man-baby to the US Presidency, if not more. I think we are aware of this issue.

  • How do we account for unexpected consequences of other Countries decisions on Australia, if indeed we can?

What would a globalised democratic system of organisation look like?

It is all quite interesting.

I imagine something similar to the court system. People like Judges with a deep knowledge of the subject interacting with a Jury (random selection of citizens) to make decisions.

nope, that’s the problem examined in the link @Frew posted. No one-person can comprehend or manage the increasing complexity of human society. Then to press-gang muggles to make decisions is madness …

I’d go with Noocracy.

I don’t understand your argument for how what I said is complex. You take a similar system applied to justice (Judges/Lawyers/etc bring the technical expertise, Jurys bring the perspective of the electorate) and split that off into every discipline of governance under the sun and there you go.

I think I knee-jerked on “random selection”. I’m saying, understanding and making good decisions in an increasingly complex human society is beyond the ability of a few people. In fact Yaneer Bar-Yam (Director: New England Complex Systems Institute) agrees with you …

“We’ve become fundamentally confused about what the decisions are, and what their consequences are. And we can’t make a connection between them,” he added. “And that’s true about everybody, as well as about the decision-makers, the policymaker. They don’t know what the effects will be of the decisions that they’re making.”

Bar-Yam proposes a more laterally-organized system of governance in which tons of small teams specialize in certain policies, and then those teams work together to ultimately make decisions.

“We end up with people who will say, ‘I will do this, and things will be better.’ And another person who will say, ‘I will do this. And things will do better.’ And we can’t tell,” he said. “Right now the danger is that we will choose strategies that will really cause a lot of destruction, before we’ve created the ability to make better decisions.”

Perhaps my thought might be better posed as a question: What duties of a politician could be replaced with technology?

Understandable but the idea that we can have meaningful elections for thousands on thousands of positions of power is ridiculous. (In fact I would argue that a vote for any representative rather than policies isn’t meaningful.) At a certain point the engagement of the electorate is even less informed than a random number because shoving a name/label in front of their face repeatedly will replace the role of knowing anything about the applicants.
Sortition is as old as Democracy.

Lawyers have AI helping them sort through documents now. Doctors have AI helping with diagnosis.
I’m sure similar projects are happening for politics… hopefully in governance too, instead of just the fucking elections.


That is exactly what this thread is about. :slight_smile:

@AndrewDowning put forward an idea for a deliberative democratic platform called Polly, which was informed by the failures encountered by the Pirate Parties who instituted liquid democracy.

The original concept is outlined in this presentation which was put to the 2012 Congress. Andrew is planning to get the project going again.

I meant AI projects helping sort information for actual complex governance tasks affecting real world issues, not just coming up with more complicated voting systems.

As far as “Designing Online Democracy” I think it’s all overthinking it.
You want consensus on an issue among members? It’s the same as offline, bring together members to discuss and come to a consensus. A forum, a chatroom parliament, the kitchen of your flat are all perfectly good venues to discuss and yes/no on issues.
Any more complicated system of voting you design is without question gameable and will only encourage competition and building power structures over consensus.