Augmented Democracy

Proposal for a “decision network”, an internet-enabled meritocratic social network for collective decision-making

 

[This proposal was funded by Threshold Foundation (www.thresholdfoundation.org ) and Wallace Global Fund (www.wgf.org ), was developed and tested in early 2006. We are currently exploring collaborations with other collective-decisionmaking initiatives, and have some development funding.]

 

"Democracy is the worst form of government except all the others."    -  Winston Churchill

Until now, that is. The Internet presents fantastic opportunities for collective decision-making that can re-energize civic participation, by solving a core problem of democracy, namely that one-person/one-vote guarantees that the wisest among us will be devalued.

We propose a simple but powerful new way to make collective decisions of all kinds meritocratically, using new Internet social-networking tools. We will use it first with a stellar community of independent media influencers and their networks, for deciding how to allocate a fund of approximately $10-20,000 to projects submitted online. Each participant will be required to donate $10-50 in order to vote, which will create a second, and thus “evergreen”, fund.

The core idea is that each participant gets an equal number of votes (e.g. 10), to be exercised not by them but by their “proxies”, i.e. those they trust most to make good decisions. That simple change, from voting to delegating your vote, creates meritocracy in an equitable, natural way.

That is because the most highly respected participants get delegated the most votes. This approach combines the best of direct and representative democracy and creates a true meritocracy where merit is decided at the individual level, but aggregated for collective use.

Information technology makes this possible. New social-networking software makes it easy to allow any user to identify those they trust most, to make that information useful online (anonymously of course), and to use it to make collective decisions. 

This system is highly scaleable (see http://www.sonic.net/~daniel/pubs/steinbock-ssdm.pdf ), to literally millions of people. Other powerful near- and medium-term uses of this network are: a. the filtering of news and information, by treating the network as a large-scale “editorial board” where everyone has a voice commensurate with how much others want to hear it; b. as a “web service” that any large informal group (e.g. the World Social Forum) can use to make collective decisions without bogging down on “voting on how to vote”;and  c. revolutionizing the primary election process by allowing engaged citizens to issue an “open call” to potential candidates, and efficiently select the candidate that best fits them collectively and thus has the best chance to win.

One important outcome we hope for is that such a system could motivate civic engagement globally, by being a highly useful way for the myriad committed global citizens to productively collaborate at whatever level their time allows. There currently is no real context for such large-scale collaboration that has concrete impact.

Proposal

This grant seeks to fund the tools for building and using such a “decision network”, starting with independent media  leaders and their networks, applying it first to a collective decision on how to allocate a pool of money. We have received a grant from Threshold Foundation to provide that pool and contribute to development of the tools. We would require a matching donation of $10-50 from each participant (call them “dues”), which will double the size of the fund, or, if we choose, will create  a second fund, to be allocated in a later round. This would create an “evergreen” funding system that would create a new fund each time the current fund is allocated, and that would grow with the network.

We seek $15,000 to implement the system, and to apply it to the Threshold grant and/or to the funds raised from participant dues. This project is an initiative of Media Venture Collective, a non-profit project fiscally sponsored by the Rudolf Steiner Foundation.


How

We will tart with an initial kernel of 100 or so international influencers, who presumably have their own networks of people they respect. These would be chosen in collaboration with the grantors.

Each will be given 10 proxies to give to others, who may or may not be in the network already. Each new addition gets 10 of their own to give away. We have a lightweight and simple email process for proxy allocation.

The invitation process will be closed at some arbitrary limit, either time or total number of participants. At that point, all participants will have a precise number of votes equal to the number of invitations they have received. Late invitees will have to choose their proxies from existing participants.

Candidate projects will be submitted online (see below), possibly by invitation only from the network or the initial kernel.

Voting will take place over a designated period, probably two weeks, possibly with one or more non-binding iterations for learning purposes.

Those who don’t vote automatically pass on their votes to their proxies.

The $10 participant matching donations will double the grant fund, or be pooled into a second fund.

For example, here is a possible graph of such a network involving partipants “m”, “l”, “g”, “k”, “e”, and “n”, choosing 2, 3, 4, 4, 2, and 3 proxies respectively (the arrows). For instance, “m” and “n” each chose the other, among others, as one of their proxies. “g” chose “m”, “l”, “k”, and “e”, and the latter two chose her. The size of each circle represents their influence, and is a function of the incoming arrows. If “k” doesn’t vote, her influence moves to “g”, “n”, “m”, and “l”.  If only two vote, they still represent the rest of the network meritocratically. This can scale to much larger networks, millions in fact. (see www.arxiv.org/abs/cs.CY/0412047)

 

Who

Such a network needs to start somewhere, with at least one person, who then names 10 others that they respect, who then name 10 others, etc. The starting number is more or less arbitrary; we’re aiming for around 100 because it can be very inclusive and can accelerate adoption faster than a small number.

Let us call the first invitees the “kernel” of the network. They, and their choices of proxies, will set the “initial conditions”, the personality and values, of the network. However, we believe in general that the ultimate shape of the network is not highly dependent on the specific individuals in the first round, because anyone left out that might be included would inevitably be included by one of those in the kernel, or by someone they choose.

We expect strong participation from the first 100, and thus think that a low estimate of 1000 participants is very achievable. Depending on enthusiasm for the idea, we believe 10,000 is certainly achievable under the right conditions. We may decide to put an upper limit on the number of participants (e.g. 2000), to keep it focused on true experts and to minimize variables.

One important factor affecting how the network grows will be the size of the fund it will be allocating. It’s pretty clear that it will be easier to get people to participate the larger it is. We are working to maximize that through the matching donations of the participants, and through comparable grants from other small foundations

Project Submission and Review

The process of project submission is relatively straightforward. At the very least, we plan to start with an existing system provided by Planetwork, which can be seen at http://www.planetwork.net/2004conf/interactive/index.html . It supports most of the necessary features, and we have the code so can easily customize it.

Submission

The system will support a full range of choices for where projects come from, from fully open, where anyone can submit, to open only to participants, to open only to the kernel, to centrally curated by us.

For the purposes of this project, we plan to keep it simple and limited. We propose to solicit projects from the kernel and their proxies only, but reserve the right to disqualify any project for any reason, to avoid distractions if necessary. So the process would be:

a.                 Each member of the kernel and their proxies can suggest one or more projects for consideration, whether their own, or one they know of and like.

b.                 The “owner” of that project will then be given a password to use the project submission system.

c.                 Upon our review and approval, the new proposal will be made “live”. We don’t expect to deny approval in very many cases, as we want the system to decide what’s appropriate.

Criteria

An initial pass at proposed criteria is:

1.       Media focus – Media must be a primary, not secondary, component

2.       Public interest – Project should benefit the public significantly, not necessarily to the exclusion of private beneficiaries.

3.       Sustainable business model – Project should have a business model that provides a credible means of long-term viability

4.       Scaleability – Project should be able to benefit large user bases, even if it only effects small numbers initially

5.       Tools and infrastructure, not content – Projects that provide widely useable tools and infrastructure (e.g. publishing and distribution) are considered preferable to content projects (e.g. the production of a documentary). Pure content projects will be disqualified unless there is some compelling reason not to.

6.       Amount of funding – Current thinking is that all should be targeted at the full amount of the fund, to make comparison straightforward.

Conflict of interest

With participants both proposing and approving projects, there is of course the potential for conflict of interest.  We believe that is a necessary challenge built into this concept, because we want influential people both for their help in making decisions and for their knowledge of projects potentially of interest to the whole network.

We believe that conflict of interest problems can be minimal, for three reasons:

1.       Voluntary recusal – Participants will be told that they are not to vote for any project that they have a direct interest in.

2.       Enforced recusal – In most cases, we can detect conflicts and any participant that is the submitter of a project or clearly associated with it can be flagged as ineligible to vote for it

3.       We believe that 1 and 2 might not even be necessary, because it is highly unlikely that any one participant will have enough influence to get a project approved without major support from other camps on its own merits.  Because it is all automated, we can actually look at the results both ways, and decide  if it makes enough of a difference to enforce.

Rating and Voting

One interesting aspect of this process is that voting can be viewed simply as a final version of an on-going rating process. Each participant can rate projects at any time, and can see the current up-to-the-minute tabulation of all ratings. This will be a useful tool for busy participants in deciding which of the many projects they want to spend time reviewing. There is a danger in this, in that the projects that get early encouragement will tend to get reinforced, sometimes unfairly. This can be mitigated by not making the rating public until all submissions are in, and by requiring participants to each review at least one random project, akin to peer-review of scientific papers.

Comments on each project will be easily added and publicly available.

Voting then becomes simply assessing the current state of ratings at a chosen time. This will in fact be more formal, in the sense that all participants will be notified that a vote is in progress and that they have a deadline. We will probably hide the current state of ratings, but not the comments, when calling for a vote so as not to unduly influence the outcome.

 

Next phase uses

In future versions, it would be applied to many other “collective intelligence” applications, such as:

A filter for news and information .- How do we decide what to pay attention to? Usually, we do it through recommendations, from friends, experts, or “the public”. This system is ideal for such uses, because it integrates popularity and expert opinion in a seamless spectrum that includes everyone but avoids the tyranny of the majority. Each participant can highlight stories, and influence their prominence in direct relation to the respect others have for their opinion..

It is ideal for a “people’s news hour”, which could be a serious alternative to the choices made by corporate media. Participants could choose the stories they think need to be covered, and a daily news show could be built around those. Because the most respected have the most influence, the choices made could easily be argued to be more trustworthy than any centrally-controlled editorial board, and would give a real incentive for participation.

A radical new kind of election primary – This approach could completely replace the Party-centric, back-room, money-dominated process of selecting a candidate (e.g. for a House seat in Congress). It would put citizens in charge of choosing the candidate best suited to them, regardless of how much money they have to spend getting name recognition. Here’s a possible scenario:

a.       MoveOn (or other progressive influencers e.g. Alternet, The Nation, etc.) picks a Congressional district to try it out on, let’s say one in upstate NY.

b.       They email all their members in that district and invite them to participate.

c.        Those participants invite their proxies, other voters in the district whose vote they trust, who also continue the process, presumably reaching every citizen that actually cares.

d.       They issue an open call to candidates to submit their qualifications to an online review system.

e.        Each participant rates the candidates, but ONLY if they have real knowledge of them, good or bad

f.        Since non-voters’ influence cascades to their proxies, those who know the candidates well are voting not just for themselves but for their friends, and their friends, who don’t.

The results of this are :

a.       Any candidate can test the waters for zero cost

b.       Highly qualified candidates can get traction early regardless of money.

c.        Time spent fundraising can be reduced because potential funders can look at the results and judge for themselves.

d.       Candidacies based on many small citizen donations will be encouraged because the candidates will already have a relationship with a motivate subset of their constituency.

e.        This system could also include a commitment from the participants to contribute $50 or so to the winning candidate, giving them a running start

f.        Unqualified but rich candidates will have a much harder time “buying” the job because they’ll be fighting bad results from the primary.

 “Domain-specific” uses – Person-to-person proxy assignment is highly dependent on the domain of expertise. Someone I respect for financial decisions might have terrible taste in music, in my opinion. Future versions will allow each proxy assignment to be domain-specific  (e.g. music, news, food, finance, human rights, technology, etc.), allowing the expertise of the participants to be very focused on specific areas.

Web service for collective decision-making – A natural evolution for this system is to offer it as a decision-making tool for any ad hoc group. For instance, the World Social Forum might query the system for the influence that each of its attendees has in a particular domain (e.g. human rights, or water), then invite them all to vote on a decision around that issue. Those who don’t vote get represented by their proxies, or proxies of proxies, who do.

 

Problems we solve (opportunities we exploit):

Problem/Opportunity 1 – More engaged participation: 

“One-person/one-vote” means that the wisest elder is no more influential than the least informed.  Such a meritocracy motivates our most valued citizens by giving them a sense that their expertise matters.

Problem/Opportunity 2 – Better decisions

By definition, if our wisest are devalued, our collective decisions are unlikely to be the best.

Problem/Opportunity 3 – Constructive financial mobilization

MoveOn, the Dean campaign, the tsunami, and others have proven that large constituencies will mobilize small donations on a large scale, given a good reason. So far, all such reasons have been reactive and/or defensive. We have a great opportunity to change that, to mobilize financially in a constructive way. That is the core of the vision behind Media Venture Collective, to mobilize citizen financing of a media revolution the benefits the planet.

The challenge of that approach is that people respond to urgency; it’s difficult to motivate them around long-term opportunities. Thinking big is required, but unrealistic at the same time.

This project can change that, by priming the pump. The $10-15,000 grant fund from Threshold provides the motivation to participate, and the $10 participation contribution doubles that motivation. Presumably, the network will grow with each round, and participants may give more than $10, so the fund will grow.

Problem/ Opportunity 4 – Too much information

Such a meritocracy is the ultimate tool for rating information, because it takes the quality of the recommender into account, aggregates ratings from large numbers of people, and can be tuned to the individual. See above under Next Phase Uses

Problem/Opportunity 5 – Global collaboration

Global civil society is rich with great people, the so-called “second superpower”  (http://cyber.law.harvard.edu/people/jmoore/secondsuperpower.html ) . However, in a world dominated by winner-take-all politics, that superpower has no functional context within which it can operate, and thus is unable to focus its power to set the social agenda, make decisions, and have real effects. This approach has the potential to change that.

Goals

1.       Powerful network of influencers – It is hard to over-estimate the value of a social network of the most highly respected influencers in global civil society and what they think of each other. This has many powerful uses, and creates a simple context within which that global community can collaborate at whatever level makes sense to them. Done right, it would be the only such network necessary.

2.       Evergreen funding mechanism – We hope to use the initial grant fund to create the funding for the next cycle through the small matching grants of the participants, and to have it grow as the network grows.

3.       The ultimate information filter – We believe that this is how humans will filter information eventually, because it automates what we already do, pay attention to trusted recommendations, and scales it up.

4.       Setting the stage for use in election primaries.

Specific Objectives

1.       2,000 great participants – We will start with 100 well-known influencers, and hopefully get them to invite 10 others each, who will do the same. Assuming a 20% participation rate, that yields 2,000.

2.       Quality use of funds – We hope that the choices of projects to fund will be high quality, and that the network will make good choices..

3.       Second fund – If objective 1 is met, and each contributes their matching donation of $10, we will have $20,000 to double the fund, or to repeat this process with.

 

Schedule and Budget

The work of this project will begin as soon as funding is available. Implementation of the network propagation and project submission systems will take two to three months. We hope to roll out September or October 2005. Our budget for the development is $15,000. The grant fund will be at least $10,000, hopefully more than doubled by participant donations, as well as potential matching grants from other foundations.

Contact

Brad deGraf

180 Marguerite Avenue, Mill Valley, CA 94941

415-847-8097 mobile

415-388-6478 office

bdegraf@pacbell.net