Brewhouse of Berliner Kindl Brewery Neuölln in 2010. Wikicommons/Daniel Markovics. Some rights reserved.

The Politics of Code is an interview series curated by Marco Deseriis with software engineers and political activists about a new generation of decision-making software that allow movements, organizations, and parties to make collective decisions online. Rather than reducing technology to the status of a tool, the interviews explore the different conceptions of political participation and democracy embedded in each software, the relationship between online and offline deliberation, as well as questions of authentication, verifiability, ownership, trust, and leadership.

Founded in 2009 in Berlin, the Liquid Democracy association has been developing an impressive array of open source software tools that support civic and political participation in youth projects, urban planning projects, participatory budgeting, NGOs, political parties, and institutions.

The best known of these tools is Adhocracy, a modular decision-making platform that allows participants to collect ideas, discuss them, and refine them in text propositions that can be further amended. This modular structure has been mostly used in civic participation projects in Berlin, but also in political parties such as the Green Party, the SPD, and institutional contexts such at the German Federal Parliament.

Thanks to Adhocracy’s popularity the association became in a few years a non-profit organisation that now employs full time twenty people working on a variety of projects. These include the third version of Adhocracy and platforms for youth participation such as Opin.me and Aula. Further, the association is currently managing a multi-purpose participation portal for the State of Berlin, which allows Berliners to take part in participatory budgeting projects, zoning decisions, environmental initiatives, and allocation of funds in their neighbourhood.

A few weeks ago, I paid a visit to the new headquarters of the association located in the former Kindl brewery in Berlin’s Neukölln district, a massive red-brick building hosting the new Kindl Center of Contemporary Art. In the spacious and still largely unfurnished rooms on the second floor, I met Rouven Brües, a PhD candidate in politics at Goldsmiths, and currently one of two managing directors of the company. Our conversation touched on a number of topics, including different implementations of Liquid Democracy (also known as delegative democracy), the relation between online and offline deliberation, participation and representation, as well as the cultural impact of new technologies on the democratic imagination.

Brewhouse of Berliner Kindl Brewery Neuölln in its former glory. Public domain.

Marco Deseriis (MD): Can you recount how the Adhocracy project came about? 

Rouven Brües (RB): Around 2008, a city-wide discussion began in Berlin about the concept of Liquid Democracy in the homonymous group founded and organised by Daniel Reichert and Frederik Wegener. Many people were interested in discussing how technology can be used to improve and change our democracy, especially in the context of the newly founded German Pirate Party. Beginning from this discussion, multiple organizations were founded. The Public Software group decided to focus on participation in political parties and developed LiquidFeedback. The founders of our organization, the Liquid Democracy e.V. had the vision that participation via delegated voting should be extended to the whole of society. So we had this very broad vision of developing participation within youth organizations, universities, political parties, cities, communes, and many other organizations. This was the higher-level vision. We initially developed a basic set of features, which are still present in the more recent versions of Adhocracy. This is because we realized that many of these participation processes boil down to certain steps. We also learned that it is very challenging to connect online and offline participation, and that there is much more to a decision than just the final vote. Since the beginning we focused more on deliberation and deliberative notions of democracy.

MD:  What is more important than the final decision? 

RB: The process itself. We are particularly interested in making the whole decision-making process – from discussing a topic to preparing a decision to the actual decision – more democratic and transparent. So since the beginning we focused more on deliberation and deliberative notions of democracy. This entails asking questions such as “How can you engage people and inform them? How can people interact with each other and discuss their interests before they come to a final conclusion or decision?” We decided to break down this process in modules. These include collecting ideas, setting agendas, discussing results, georeferencing ideas on a map, and cyclical events. Because every participation process is slightly different, not all modules are always used.

MD. What is a cyclical event? 

RB: Every group and organization tends to hold meetings and organize recurring events. We have designed a tool that can help participants set the agenda so that the most important topics are always discussed first and the ones that cannot be discussed are automatically put on the agenda of the next meeting.  

MD: Can you explain how the relationship between online and offline deliberation typically unfolds in Adhocracy?

RB: As I said, every user case is slightly different. I can mention the example of the former Tempelhof airport, which was converted into a city park in 2009. In 2014, Berliners decided in a referendum to keep the airport field free of construction. A law was passed that stated that the citizens had to be consulted on how to design the new park. In November 2014, we launched an online platform to collect ideas from Berlin residents. Ideas ranged from the kind of sports that could be performed there to how to valorize Tempelhof as a natural reservoir and historic site. Subsequently, these ideas were discussed, clustered, and prioritized in a series of offline workshops, which were open to everyone, regardless of whether you had contributed through the website or not.

For example, somebody proposed on the website to keep the park gates open at night. Because there was so much interest in this topic, we held a special workshop at the park every other week to discuss it. The results of this discussion were uploaded as “protocols” to the website. These protocols were also participatory, i.e. users made amendments to the draft of the final text, which was eventually included in the Development and Care Plan for the park. This final text was approved by the Berlin Parliament in April 2016.

And even after the approval, the process is not finished. The plan states that people can meet at the field twice a year and participate in a forum that has the power to determine how the park will be used the following year. At the first forum, people could elect representatives in the governing board of the park, and will be able to do so once a year. The whole process was very constructive, there were moments of discussion, but there was never a moment in which the City thought about backing down. Citizens and the administration engaged in what is sometimes referred to as a “co-creation” process.

Screenshot of Tempelhof project.

MD: Was the website built by Adhocracy?

RB: Yes, it was first built with Adhocracy 2 and then with Adhocracy 3.

MD: What is the difference between version 2 and version 3?

RB: Adhocracy 3 uses the same framework of Adhocracy 2 but it is a re-write. The most significant difference between the two is that in the case of Adhocracy 2 those who were interested in using it had to come to our platform, set up the project, and get all the participants to sign up. With Adhocracy 3 we have designed a process that everyone can use and embed on their own website in the same way as they can embed a YouTube video. For example, if you have a political campaign, all participants can access the campaign through their own website. It doesn’t matter where people participate from, everyone participates in the same discussion, which is synchronized across the different websites that embed Adhocracy 3. On the other hand, Adhocracy 3 has design limitations for user navigation and is a very complicated development process. We wrote the whole backend, the whole architecture by ourselves, and you have to be a really knowledgeable Python programmer to be able to program this framework. This means that it’s hard for us to find new developers as they might find it difficult to get into the code. We have designed a process that everyone can use and embed on their own website in the same way as they can embed a YouTube video.

MD: Does Adhocracy allow users to delegate their vote to other users?

RB: It does, but we never used this feature. In the case of Tempelhof, for example, there was never a binding decision, it was an open discussion process. For us, it is very important to communicate in advance the various steps of the process so that participants know what to expect from it. If you have delegated voting, the assumption is that in the end there will be a final decision online, which isn’t always the case.

MD: Are you suggesting that in the case of civic participation projects, delegated voting is not necessary because people tend to meet and trust each other through local meetings?

RB: Not exactly. I still see a great value in delegated voting, it’s a really interesting concept. We didn’t forget about it and we are still trying to achieve it in spite of the fact that we never had a user request for it. After we built the first version of Adhocracy 2, we offered this functionality to the Green Party of North Rhine-Westphalia in 2014. They wanted to draft their state party program online and to make the process as transparent as possible. The delegated voting feature was active, but no one used it.

That’s why I always argue that we first need to change our political culture. Although technical possibilities influence what we can imagine democratically, nothing is going to happen unless people envision a democracy where you can delegate a vote. For now, organizations like political parties seem more interested in using this kind of tool for opening up the discussion. For example, the SPD asked their members and the interested public to give feedback on their ideas about digital change in Germany in a project called Perspectives for Germany 2025. But even in that case the emphasis was on discussion rather than decision.

MD: Does this mean that Adhocracy allows for a discussion phase without necessarily leading to a final decision?

RB: Yes, the second version was not divided into phases. You could just activate the different tools at different times. We also suggested making the time frames very clear in the beginning. In the new versions of Adhocracy 3, Opin and Aula these different phases are still optional but more streamlined in the software.

Tempelhof project.

MD: Can you describe the different phases in a typical participation process?

RB: The “standard” version of the process begins with an announcement phase where you announce who is going to participate, how long will it take, what is the outcome, whether there are offline meetings, and so forth.

The importance of letting people know what they can do and how they can engage cannot be overstated as online tools do not set the conditions for participation by themselves. Then there is a second phase where topics, ideas, and proposals are collected both online and offline. For example meetings that might be held in different neighborhoods lead to the formulation of ideas that are then uploaded and further discussed online. In the third phase, the results of the discussion are incorporated in a text proposition that can be further amended. Finally, people can take a vote but, as I said, this is optional.

MD: Why is it optional?

RB: Because it’s a question of representation, or lack thereof. To begin with, you cannot guarantee that the people taking part in a participatory process represent the whole population. For example, if an administration does not advertise the process it can never claim it’s inclusive as there will always be people who have never heard about it. The same goes for voting, you can never guarantee that everyone who has a right to vote can actually votes. So we argue that you have to advertise these participatory processes and establish them in the long run as possibilities to engage. For us, it is already a democratic achievement if you have the possibility to participate. That’s perhaps our strongest normative claim.

MD: Why is it so important to emphasize the possibility to engage?

RB: Because we meet a lot of skeptics. Let’s take again the example of Tempelhof airport, which in our view was one of our most successful participatory projects. Although there are nine thousand people registered on the website, Berlin has more than five million residents. So the City has good reasons to wonder whether this process is really representative. We respond to these objections by arguing that there is a ladder of participation. Registering on the website is not the first step of the ladder. The first step is to click on the website to figure out what is going on. Over 120,000 people visited the website from November 2014 until today. Out of these 120,000, in the first participation phase there were 14,000 registered users, 354 proposals, 1736 comments, and several thousand preferences were expressed for one of the proposals.

Obviously to “like” something takes much less effort than to write a proposal. That’s why we aim at the quality of participation and not necessarily representation. For these participatory processes to become really representative administrations will have to make a regular use of online tools and regularly publicize them. Now every district that plans civic engagement knows that it has online tools at its disposal and that it can use them to engage the citizens. Only when this becomes the norm, will we be able to claim that these participatory processes are also representative. Now every district that plans civic engagement knows that it has online tools at its disposal and that it can use them to engage the citizens.

MD: One could also make the claim that the ones who engaged in the process are more likely to use the park. So perhaps there is no need for them to really represent the orientation of five million Berliners, many of whom have never been to Tempelhof, or will just be happy with whatever decision is made because for them the stakes are really low.

RB: Yes, we always have to consider the output of a decision. After all, the connection between the output of a decision and its effects on my personal life might be rather small. That may explain why “only” nine thousand people registered on the website. If there was a binding decision on how to allocate the budget of the State of Berlin, I promise you that more than nine thousand people would participate because it would affect their personal life on a deeper level.

MD: Let’s talk about authentication. Are the platform users required to use their legal name or can they register using a pseudonym?

RB: They can use a pseudonym if they want to.

MD: Is the pseudonym what appears as their screen name?

RB: No, you can use a pseudonym when you register on the platform. For the Tempelhof project, the city was fine with that. Currently, the State of Berlin is planning to create a service account that identifies individual citizens and companies alike, with which they can engage with the administration. There is a plan to link this account with the participation platform. But we would design the platform in a way that the knowledge of who wrote what is still under the control of the individual. The platform will know that you are an authenticated user via the service account, but it will not know who you are. Conversely, the service account will not know what you write. My feeling is that the administration is very happy with the participation process as it is, because they use the data for their work. For them the more committed individuals participate the better, it doesn’t matter whether they are using their real name or not. Of course, if the platform was to be used for online voting, it would be a whole different story.

MD: Speaking of the impact of Adhocracy on institutional decision-making, can you tell me a little bit about the adoption of  the platform at the Bundestag?

RB: The German Federal Parliament has research commissions, which study a variety of topics. From 2010 to 2013, the Parliament instituted the Inquiry Commission “Internet and Digital Society.” The commission had seventeen members (which were allocated according to the percentage of each party at the Bundestag) and studied how technology impacts on German society.

The commission was organized in twelve groups which were tasked to draft a final research document and make policy recommendations to the Parliament. Each of the groups used the Adhocracy platform to a different extent, but the one that used it the most was the group that focused on political participation, e-government, e-administration, and so forth. At a certain point, someone in the commission suggested to add an eighteenth member, which would be the public, and which would interact with the other members of the commission via Adhocracy. After some struggle, the commission agreed to host the platform on a separate server from those of the official functions of the Bundestag. There, people could read about the groups, make propositions, and even make direct contributions to the text. The public’s suggestions were integrated, sometimes literally, in the final text that was eventually voted on by the commission. This was a big success for us as it allowed the public to give direct input on proposed legislation, something that would have not been possible without Adhocracy.

MD: I’ve heard that one of the hurdles was the problem of authentication as there was no certainty that the people who were contributing ideas were entitled to do so. How did the research group address this issue?

RB: Since the whole process was consultative and non-binding, no hard authentification was needed as everyone was entitled to take part.

Often, when we act as consultants for cities or other organizations we encounter a mild form of skepticism – the conviction that individuals who should not take part in something will somehow get involved. After all, why should someone living in Cologne have a say in a construction project in Berlin? To this kind of objection, I usually reply by inviting a change in perspective. The problem is not the Cologne-based architect who may make a brilliant contribution to a construction project in Berlin. The real challenge is to get committed people to spend their time, knowledge, and effort on a problem to try and help solve it. The real challenge with online consultation and participation processes is to motivate people to participate and not to limit the participants on the basis of their geo-location or other restrictive criteria. Unless you want to have a binding online vote, which, as I said, has a completely different set of implications. The real challenge is to get committed people to spend their time, knowledge, and effort on a problem to try and help solve it.

Liquid Democracy at work.

MD: Is the blockchain technology a possible solution to the problem of authentication for online voting?

RB: Yes, the blockchain technology is so hyped these days because it is seen as a way of solving this conundrum. There are different organizations like the Ethereum Project and Democracy Earth that are working on it. Blockchain can guarantee that eligible people are authenticated, but the vote remains anonymous. Everyone can see the results, verify their integrity, but you do not know who voted and how.

The blockchain mining process requires a lot of computational power and it is still very expensive to solve these cryptic puzzles, but the promise is there. Again, this is where I argue that these technological affordances change the way we think about democracy. If we know that secret digital voting will be safe, then we are confronted with a whole new way of imagining democratic life. Such knowledge might change democratic processes even before the technology is implemented.  

MD: My final question is more theoretical. I was wondering whether you would agree with those who argue that by facilitating transparency these participatory tools are turning our political orientation into another precious “value” or indicator for the data mining industry.

For techno-skeptics such as Jodi Dean and Evgeny Morozov, open source software, civic technologies and the whole fetishization of digital transparency and networked participation only work in the service of informational capitalism…

RB: It is important to keep in mind this critique as a way of analyzing the work that we do. There is no doubt that since its inception the Liquid Democracy association has undergone a process of professionalization. This means that we are not only developing the software according to the vision that we have but we are depending on financial means and investments of private clients and EU research projects to do so.

These clients and projects have their own ideas and needs, which have to be incorporated into the process of developing the software. So you could say that we have become alienated to a certain extent from our original intention of building a radically liquid democracy.

Yet I am also really optimistic about changing our democratic culture. And here we have found the way of doing it, albeit on a certain level. We may not be transforming democracy at the root yet – it’s not completely radical because we do not hold actual electronic elections, we do not control the count, and we cannot change the whole democratic system through these online tools. But we are slowly carving a way into the democratic fabric. We are now taking the first steps through which citizens, the body politick, the administration may change, to make these processes more open and more accessible to everybody.