En attendant une hypothétique invasion extraterrestre, il y en a une en cours, très concrète : les comptes spam pullulent en effet sur les plateformes de participation en ligne. Si d’autres types de plateformes sont également touchés (coucou 👋 Elon), les civic ou govtech le sont particulièrement, les moteurs de recherche appréciant beaucoup les plateformes institutionnelles… Explications.
De quoi parle-t-on ?
Les plateformes de participation citoyenne en ligne mises en place par des acteurs publics sont de plus en plus la cible d’acteurs malintentionnés qui créent des comptes spams et publient parfois des messages indésirables sur leur plateforme. Dans cet article, nous revenons sur les origines de ce problème malheureusement bien courant sur internet et sur les solutions que nous avons développées pour y répondre.
Un compte spam typique
Un commentaire indésirable
Pour quoi faire ?
Les acteurs en question utilisent des programmes ou des services automatisés pour créer des liens vers leur site web pour en optimiser le référencement naturel. En effet, plus un site reçoit de liens entrants venant de domaines à forte autorité, plus il apparaît haut dans les résultats de recherche
En automatisant la création de liens vers leur site, il remonte dans les résultats de recherche.
Il s’agit d’une pratique très répandue chez les adeptes du “Black Hat SEO”. Le “Black Hat SEO” fait référence à un ensemble de procédés et d’automatisations dont le but est d’exploiter les failles des algorithmes de référencement afin de manipuler les résultats de recherches. Ces techniques peuvent s’avérer très efficaces à court terme jusqu’à ce que les algorithmes soient mis à jour. Elles sont en général pratiquées par des entreprises spécialisées, peu recommandables, agissant à la frontière de la légalité.
Un phénomène répandu
Sur Internet
Les comptes spams ne sont pas un problème nouveau sur internet. Ils sont d’abord apparus avec l’adoption massive des emails puis se sont vite propagés sur les réseaux sociaux. Régulièrement des études montrent qu’un pourcentage significatif de l’utilisation d’internet est lié aux activités de spam (60% des emails, 8% des comptes instagram en 2015 etc.).
Il n’est pas étonnant que tout une industrie se consacre à proposer des solutions à ce problème. Le géant américain Cloudflare, qui offre une large palette de solutions visant à sécuriser les sites web, est évalué à plus de 35 milliards de dollars et gère plus de 10% du trafic mondial devenant ainsi une mesure approximative de l’activité des spams sur internet.
Si aujourd’hui votre boîte de réception email n’est pas remplie de spam, c’est essentiellement parce que toute une batterie de brillants ingénieurs ont développé des algorithmes suffisamment sophistiqués pour trier automatiquement le flux de mails entrants que vous recevez.
Sur les plateformes de participation citoyenne
Ces dernières années, un nombre croissant de collectivités et d’institutions ont ouvert des plateformes participatives. Après quelques années d’exploitation, elles sont devenus des cibles de choix pour ceux qui pratiquent le “Black Hat SEO” pour trois raisons :
Les noms de domaine qu’elles utilisent ont une forte notoriété (ex : senat.fr, paris.fr) et permettent de remonter rapidement dans les résultats ;
Ce sont des sites webs dynamiques qui perdurent dans le temps ;
Il est possible sans restriction de se créer un compte et de publier des messages.
Les différents éditeurs ont beau chercher à se protéger avec différentes solutions comme les captcha, rien n’y fait : les comptes spam trouvent toujours le moyen de se créer un compte.
Ci-dessous, quelques échantillons tirés des principaux éditeurs de plateformes participatives. A titre d’information, nous ne vous montrons ici que les exemples les plus politiquement corrects.
Le problème est si répandu que la plupart des éditeurs font le choix de brider les fonctionnalités de recherche sur leur plateforme pour éviter qu’ils soient trop facilement découvrables. En effet, sur Cap Collectif et Citizen Lab les barres de recherche permettant de rechercher des utilisateurs sont le plus souvent désactivées. D’autres font carrément le choix de ne pas proposer de profil utilisateurs publics ou de les rendre non cliquables.
Ce n’est pas une raison pour paniquer
Chez Open Source Politics, nous avons choisi de ne pas brider les fonctionnalités de recherche du logiciel libre Decidim. Nous trouvons regrettable de limiter les possibilités de découverte et d’exploration des usagers de nos plateformes.
Afin de rassurer nos clients, nous avons progressivement constitué une FAQ.
Le problème porte-t-il préjudice à la participation sur la plateforme ?
Non, la participation n’est pas impactée par la création de ces comptes. Par définition, les comptes spam sont pour la plupart des comptes inactifs, leurs créateurs n’ont aucun intérêt à participer à des démarches lancées sur la plateforme. En revanche, certains de ces comptes se permettent de déposer un ou deux commentaires à certains endroits, par exemple dans les commentaires des propositions. Ces commentaires sont immédiatement repérés et modérés. Cette pratique reste très marginale néanmoins.
Existe-t-il des risques de piratage ou de fuite de données liées à ces comptes ?
Non. Il s’agit de comptes utilisateurs standard. Une inquiétude légitime pourrait subsister concernant les utilisateurs qui pourraient cliquer sur des liens qui renvoient un site dangereux. Heureusement dans sa dernière version Decidim propose un écran intermédiaire qui avertit l’utilisateur lorsqu’il clique sur un lien externe à la plateforme.
La modale qui s’affiche au clic d’un lien externe publié par un utilisateur non admin.
Est-il possible de bloquer ces utilisateurs ?
Oui. Sur Decidim, à partir de la version stable 0.24, il est possible de bloquer un utilisateur.
Le nombre d’utilisateurs affiché sur la barre de recherche globale tient-il compte des comptes spam ? Oui, les résultats de recherche référencent tous les utilisateurs qui ont créé un compte, dont les spams. En revanche, la statistique mise en avant sur les pages d’accueil et dans les espaces de concertation ne comptabilise que les participants effectifs, qui ont réalisé au moins une action sur la plateforme. Il est préférable de s’y référer.
Comment y remédier ?
Nous ne sommes pas pour autant fatalistes, des solutions existent et permettent de traiter ce problème et nous les appliquons.
Un problème connu de la communauté Decidim
C’est un phénomène bien connu de la communauté : un fil de discussion y est consacré sur le dépôt Github, ce qui permet aux développeurs d’apporter des solutions complémentaires à ce problème protéiforme. La dernière version stable de Decidim apporte notamment une solution de long terme en ajoutant un attribut “no-referrer” sur les liens sortant afin qu’ils ne contribuent pas à améliorer le référencement des sites qui se livrent à ces pratiques.
Ce que nous faisons chez Open Source Politics
La solution proposée par Decidim rend inutile à terme les pratiques de “Black Hat SEO”. En revanche, à très court terme, nos clients sont toujours confrontés à la problématique. De nombreux comptes spam sont visibles sur leur plateforme et certains renvoient parfois vers des sites frauduleux ou affichent des photos de profil pour le moins suggestives.
Notre solution basée sur l’apprentissage automatique (machine learning)
Notre équipe technique a développé un programme qui utilise un algorithme d’apprentissage automatique et qui réalise des pondérations sur les informations du profil afin de déterminer s’il peut s’agir d’un compte spam.
Après une phase d’entraînement du modèle avec des données anonymisées, nous sommes maintenant capables d’attribuer une probabilité de spam à un compte créé sur Decidim en nous basant sur les informations renseignées et son activité.
En parallèle, nous avons développé une tâche automatisée sur Decidim qui chaque jour communique avec l’algorithme et permet automatiquement de signaler tous les comptes dont la probabilité est supérieure à 70% et de les bloquer quand elle est supérieure à 99%.
Diagramme fonctionnel de notre tâche automatisée contre les spams
De cette façon, non seulement nous empêchons les comptes spam de publier sur votre plateforme mais ils n’apparaissent plus sur votre plateforme.
L’administrateur est notifié chaque jour du nombre de comptes signalés et bloqués, il peut alors se rendre dans le back-office pour traiter les comptes signalés. Quand ils sont bloqués, les utilisateurs reçoivent une notification email. Un email de contact leur est communiqué en cas d’erreur, l’administrateur peut alors rétablir leur compte à tout moment.
Exemple d’email envoyé par la tâche automatique
Cette technologie a été déployée sur les plateformes les plus affectées par les comptes spams et nous avons pu constater des résultats plus que satisfaisants.
Elle a permis de bloquer des milliers de comptes spams qui parasitaient par leur présence et leur contributions les instances Decidim de nos clients.
Ces dernières bénéficient maintenant d’une protection permanente contre les comptes spams puisque notre tâche automatique effectue une vérification journalière.
Notre solution de captcha accessible
La tâche automatique de détection des spams règle le problème des comptes spam seulement en aval, une fois que le spam a déjà infiltré la plateforme. Pour plus d’efficacité, nous souhaitions mettre en place un CAPTCHA afin de rendre l’inscription plus difficile pour des robots ainsi traiter le problème en aval. Ne souhaitant pas utiliser la solution offerte par Google, qui en plus d’être propriétaire ne respecte pas les standards d’accessibilité, nous avons développé un captcha 100% textuel basé sur le projet open source Act as text captcha. Au moment de l’inscription les utilisateurs se voient posés une question simple pour un humain mais difficile pour un robot (ex : “Parmi les mot suivant lequel est un animal : chat, courgette, carotte, bureau ?”).
La page d’inscription de la plateforme Decidim du Département Loire Atlantique
Conclusion
Les pratiques de “Black Hat SEO” évoluent en permanence et sont toujours plus sophistiquées. Apporter des réponses à cette problématique nécessite une approche évolutive et diversifiée. C’est face à ce type de situation que nous sommes heureux d’avoir fait le choix de contribuer à un logiciel libre et open source composé d’une large communauté. Les travaux des uns profitent aux autres et permettent de répondre de manière satisfaisante aux challenges les plus ardus.
Notre approche basée sur l’apprentissage automatique ayant montré de bon résultats nous avons fait le choix de l’intégrer par défaut et sans surcoût, ainsi que notre CAPTCHA textuel, pour tous nos clients utilisant notre version générique de l’application Decidim à partir de la version 0.25. Nous allons continuer d’améliorer notre modèle, notamment grâce à un partenariat avec des étudiants de l’INSA Lyon dans le cadre du programme T4G. Comme à notre habitude, nos sources sont disponibles sous licence AGPLv3 sur Github, n’hésitez pas à venir contribuer :
In a context of institutional mistrust, many hopes are pinned on digital technology. Civic tech aims to strengthen citizen engagement and participation. However, 13 million French men and women have difficulties with the use of the Internet. This context could quickly slow down the development of civic tech. How can we guarantee the inclusion of remote audiences in the digital tools for citizen participation? This paradox was at the heart of the round table discussion "Digital inclusion - an oversight of participation" during the National Meetings on Participation in Mulhouse.
Anne-Claire Dubreuil - Digital Transformation Project Director at Sicoval
The speakers stressed the importance of associating digital mediation and citizen participation in order to reduce inequalities in access to digital technology. Indeed, online participation is meant to be free and open, but it must be accessible to as many people as possible. However, it should be noted that at present, civic tech is still far from reaching the general public and that its participants have fairly homogeneous sociological profiles. In addition, there are sociological biases that may be introduced directly into the computer code by the developers. In the IT development sector, the majority of people are still men with a high level of education and from privileged social classes. Civic tech must therefore pay close attention to this phenomenon. Otherwise, the models developed will be non-inclusive and de facto non-democratic.
Digital inclusion at the heart of the Decidim and Open Source Politics project
Since its inception, Open Source Politics has placed great importance on including a wide audience in its missions. With this in mind, our teams strive to simplify the user experience and improve the accessibility of the platform. Open Source Politics also strives to adapt the communication of its platforms to different audiences. For our partners, we produce Easy to Read (FAL) content for users.
Decidim is a project that advocates for the creation of inclusive digital spaces. Features have been designed with this in mind. For example, the represented users allow an administrator to register a contribution "as" in order to make the contributions of people who are far from digital visible on the platform.
In order to fight against the gender bias introduced directly into the code, the Decidim association has set up the DecidmFemDev Program which grants scholarships to women and non-binary people wishing to contribute to the development of Decidim.
The example of Emmaus Solidarity
The Decidim instance of Emmäus Solidarity
As an example, Open Source Politics helped EMMAÜS Solidarité to implement an internal consultation to determine the association's strategic orientations. The platform has 284 subscribers, 74 proposals and 12 meetings. In addition, EMMAUS Solidarity also launched two surveys: one for employees, volunteers and members, and one for the people it supports. In this spirit, the organisers used the " Represented User " feature during the meetings with supported persons. Our consultants participated in several of these meetings to assist in person the remote audiences who wished to contribute. The results provided a diverse input reflecting the different stakeholders of Emmäus and contributed to the development of the association's strategy for the period 2020-2025.
Digital inclusion, an eco-system of actors
10% of the French population expresses difficulties with the Internet, that's a lot!
Emma Ghariani, Co-director of the Mednum
Yes, it is! And this is why we at Open Source Politics have been members of La MedNum for several years, an SCIC in which the State is also a shareholder and which brings together inclusion professionals by co-organising Numérique en Commun[s] every year. Through our contact with this ecosystem, we learn a lot and we keep abreast of the services that are developing and that can provide solutions.
For its part, the public authorities are increasingly committed to digital inclusion. Public funds have increased tenfold in recent years, from a few hundred thousand euros to 250 million, making digital inclusion a social issue to which a large-scale response must be provided. Thanks to this, new projects are emerging and the inclusion ecosystem is developing. In particular, we can mention the #APTIC digital passes which allow the financing of training in thousands of third places in France and Aidants Connect, a State startup launched to assist people in difficulty with their online procedures. A rapid intervention kit contains numerous documents and aids to facilitate the support of people who are far from computers by digital mediation professionals.
Open Source Politics is proud to be part of this movement for an open and inclusive democracy.
The Open Source Politics team wishes Decidim a happy birthday! It's impressive how far this community has come in such a short time. Let's take a look back at our meeting with this unusual digital community.
Decidim is a technopolitical project to deepen and transform participatory democracy through the design and implementation of a digital platform.
Decidim is a project based on the principles of equality, transparency, traceability and data integrity, with ethics at the heart of its approach. If you haven't already done so, read the social contract.
Decidim provides the democratic guarantees that are essential to any #civictech project thanks to the main principles that are at the basis of its architecture
1️⃣ Open to collaboration
2️⃣ Transparency, traceability, integrity
3️⃣ Democratic quality guarantees
4️⃣ Privacy and Security
Decidim is a democratically governed digital commons and an international network:
🔧 innovative companies, such as Open Source Politics, contributing to the source code and supporting the deployment and use of the tool
Decidim is fully in line with the objectives and values that Open Source Politics set itself when it was created in 2016. When Virgile Deville, our product manager, first heard about it at a Smart City conference organised by the city of Lyon and Le Monde, it was immediately obvious.
The use case on which the software was launched 4 years ago was the final convincing factor. Decidim was indeed launched to serve the co-construction of the municipal action plan of the city of Barcelona which would be used for the entire mandate. More than 40,000 citizens participated in this massive process. This participatory approach remains one of the most striking democratic innovations in the civic tech sector.
Adoption by our team
Shortly afterwards we decided to concentrate all our efforts on Decidim in :
Today we are happy to see that Decidim is becoming the standard for the implementation of digital participatory processes in France, Europe and internationally. It is used by institutions of all sizes, from local authorities to the European Commission, including the Senate and theNational Assembly.
The coronavirus crisis should be a catalyst for institutionalising the use of digital tools in parliament
This article is a translation by Open Source Politics of the article published on the Medium "Participo", an OECD publication. To read the original article by Paula Forteza, click here.
Since the beginning of the coronavirus crisis, legislative processes have stalled due to physical distance. Many legislators are now testing technologies that will allow their democratic institutions to meet, deliberate and vote despite these restrictive measures. We need to use this momentum to be ambitious in terms of participatory and collaborative legislative processes. How can we do this? By institutionalising the use of digital tools in Parliament with the necessary security and privacy safeguards.
Digital technologies to ensure democratic continuity during the crisis
The National Assembly has introduced temporary solutions to preserve parliamentary debate by using video conferencing applications for committee deliberations. However, legislative voting is still done in person and is only possible for a very limited number of members. There are both historical and technical reasons for keeping the institutions functioning mainly physically: sincerity of voting, security, tradition, etc.
Our parliament is lagging behind in modernising its processes, unlike, for example, the parliaments ofLatin America, which are pioneers in this field. In addition to using video conferencing applications, the parliaments of Brazil, Chile and Ecuador have developed their own online platforms and solutions for recording attendance, verifying quorum and voting. In at least six other South American countries - Argentina, Colombia, Costa Rica, Honduras, Jamaica and Mexico - legislative bodies have begun to experiment with virtual participation, applying it to non-decision-making spaces such as working groups or committee meetings. In Europe, countries such as the United Kingdom andSpain are moving faster than we are. We should follow these examples.
In France, one of the most common arguments against the digitisation of parliament is that it will weaken the ceremonial aspect of parliamentary deliberation. It is argued that digitisation could threaten the symbolic duties of MPs and their political weight vis-à-vis other branches of power, including the executive. Another reason may be the lack of digital literacy and skills in the French parliament. A recent study showed that only 5.37% of MPs considered themselves digital experts, 10.23% as connoisseurs and 12.65% as enthusiasts.
In order to achieve a digital parliament, technical profiles and suitable equipment should also be recruited in both chambers. As regards concerns at the political level, the doctrine needs to change. Digital parliaments may be less solemn than traditional procedures, but this horizontality is beneficial and can bring elected representatives closer to their citizens.
Data protection, security and accessibility are the conditions for a digital parliament
I am not arguing for a blind race to digitalisation: of course, there are certain limits to the introduction of digital tools in institutional decision-making. For example, Zoom has been heavily criticised for unclear data protection policies, lack of security and personal data breaches.
When considering remote voting, several security issues arise, such as fraud. The use of digital tools in Parliament facesother challenges, such as uneven internet bandwidth, technical problems (bad sound, connection problems, etc.), and uneven digital literacy and skills. However, this does not mean that we should stop modernising legislative bodies. On the contrary, we should address and solve those problems that have been made more visible through physical removal measures.
In addition to the technical challenges, the use of digital tools requires high standards of data privacy, cybersecurity and decentralisation. Firstly, by implementing the EU General Data Protection Regulation, as well as stronger safeguards against government surveillance and misuse of personal data.
In addition, free and open source web solutions should be the norm for digital tools in both institutions and public administrations. For example, decentralised peer-to-peer video conferencing applications such as Jitsi or Big Blue Button can be an excellent alternative to Zoom or Google Meet.
More broadly, civic tech can help us continue citizen deliberation in times of physical distance, provided it is open, ethical and meets a real need. As an example, I have used the Decidim platform on several occasions to establish a dialogue with and between citizens. During the COVID-19 crisis, together with 65 other MEPs, I launched the platform ' The day after', where citizens could propose, deliberate and discuss their own ideas. where citizens could propose, deliberate and vote on ideas to collectively decide on the direction to take after the crisis.
Digital technologies can support more resilient, innovative and vibrant democracies
The containment measures have shown the urgency of adapting our democratic institutions and processes to ensure their continuity, even in times of crisis. Today, almost 79% of French people surveyed have a negative feeling towards politics. Beyond the response to the emergency, such technical - and cultural - developments could contribute to strengthening citizens' confidence in elected officials and institutions by promoting participation, transparency and accountability.
There are a myriad of examples, tools and methods to support the modernisation and openness of our parliaments, for example by publishing parliamentarians' agendas and expenses, ensuring transparency of lobbying in parliament, or involving citizens in law-making.
Finally, the underlying question is not whether we need more or less digital tools in our institutions. It is about taking into account the major transformation of our society, the digital revolution, and adapting our political culture to it. We can use today's challenges to build a resilient, innovative and truly vibrant democracy. Although this discussion is about the use of digital tools, the ultimate goal is to transform and adapt our institutions to the needs and realities of the 21st century.
Paula Forteza, born on 8 August 1986 in Paris, is a French politician. She has been a Member of Parliament since June 2017, representing French citizens from Latin America and the Caribbean. She has spent more than 20 years of her life in Latin America. After several experiences in the government of the city of Buenos Aires, in the French administration, at Etalab, or in entrepreneurship, she wishes to place digital, transparency and citizen participation at the heart of the political debate in France.
The following is an edited transcript of the newDemocracy Foundation's Facilitating Public Deliberation podcast, hosted by Professor Lyn Carson, Research Director at the newDemocracy Foundation, and produced by Nivek Thompson. The interview is conducted by Graham Smith, Professor of Politics and Director of the Centre for the Study of Democracy at the University of Westminster and President of the Foundation for Democracy and Sustainable Development, on what we know and don't know about the transfer of face-to-face deliberation into an online environment.
Carson: The reason I wanted to talk to you is that we are having this discussion at the time of the Coronavirus pandemic. A lot of people have been thinking about the research questions that we need to answer if we want to think about virtual mini-publics like citizens' assemblies and citizens' juries. Questions such as: can we apply the same random selection processes if we organise a virtual mini-public? How can we ensure diversity and representativeness?
Smith: One general point I'd like to make at the outset is that I don't think there have been very good conversations in the past between people who have been involved in face-to-face deliberative processes, and people who work in civic technology and digital engagement. I think they've sometimes been in competition with each other, but they've often talked in passing. And what's been interesting to me over the last few weeks is that the containment has forced these conversations to be much more focused.
To answer your specific question about random selection, what has always interested me about the technological aspect of democratic engagement is that people who are primarily technologists have given less thought to who comes to their platform, and more to how their platform works. I don't think there's any philosophical or practical reason why you couldn't apply random selection techniques to bring people into an online space. There have been many online spaces that are closed in the sense that they are closed to a particular community.
Carson: We're going to have different skills, of course. But it's only a potential problem if we don't spend enough time with people to make them feel comfortable with the platform we're working with.
Smith: We can get people onto a platform; once they're on that platform, we have a lot of work to do, which is very different from the kind of work we would do in a room where we can see people face to face. If we use a civic lottery process to recruit people for online engagement, there are two problems. One is whether, once someone has received the invitation, they have access to the technologies and have the skills to use them confidently. And the other issue is to facilitate the space. So that we can enable the kind of inclusion that we provide in face-to-face deliberations.
Carson: Yes. I just spoke to a facilitator who did an online deliberation and another issue came up. The government, the organiser, the decision maker, may well insist on using their own platform. They often have a very cumbersome platform that they could have used to get direct input from citizens, which is just an aggregation of individual opinions, but they wouldn't necessarily have used a platform that allows for all the things that you and I might describe as deliberation.
Smith: I haven't really heard that one and I have to admit it's interesting. But in a way it reminds me of battles we've had before, when public authorities would say, "Well, why aren't my public hearings good enough? Why aren't my consultation mechanisms good enough? Maybe there's an analogy and we should step in and say, "That's not good enough for the same reasons that we said we should do citizen juries or deliberative polling rather than your previous consultation mechanism.
There are a lot of solution-isms with a lot of people saying they have the platform. It's quite dangerous when people say they have "solved" all the problems of online deliberation. "This is my application'. We have to be very careful about that.
Carson: Yes, we should remember to come back to the question of what are we trying to do here? And what would we really like people to do? There can be a path dependency.
Smith: Because with the amount of stuff that's going on in Zoom at the moment, people immediately think, "Okay, so what's the functionality? What are the possibilities of Zoom? And how can I make my process more zoomable? They feel like they're using it because they've used it before. They don't ask "Should we use Zoom? "We are in an experimental phase in terms of the platform we should use. We are also experimenting with how we should facilitate conversations on these platforms.
Carson: Yes. To me, it's just a design challenge. We've always had design challenges in deliberative democracy. So this is just another challenge, and I actually think it's quite exciting. How can we do that? How can we actually allow people to deliberate together and come to an agreement together?
Smith: It depends on the type of process you are working on. We can learn from people who are doing online education about the best methods for online learning. We could be more imaginative about some of the materials we are able to use and provide.
The UK Climate Assembly, whose last weekend was postponed, is now lined up for a series of shorter meetings. Some learning was planned for the beginning. I know that they have been broadcasting it through videos. I think the facilitator of that process was quite pleased, because she was able to say "No, you didn't do it right that time. Do it again! "So maybe you can get some of your witnesses to do better presentations.
With platforms like Zoom, people can get together in small groups with an expert. And I'm sure you've had the problem of trying to get a good expert or a good witness for a deliberative process and they can't find the time in their diary. For virtual engagement, the time commitments are less. I agree that some things are more difficult, but I'm not sure it's necessarily the learning aspect.
We worked a few years ago on asynchronous platforms. We were observing people's behaviour when we provided information and there was a dialogue in a chat room. We found that people tended not to look at the information and instead went directly to the chat room. There is a sequencing problem there.
In the kind of process we are used to, where people are selected at random, they accept because they have been invited and they consider it a special thing to do. They are willing to spend this time learning. That's one of the challenges of online spaces is that you don't necessarily know that everyone has gone through the phases in the same way that you would and will know in a face-to-face space where we can literally see what people are doing.
Carson: It seems to me that there are a lot of variations. There's synchronous where the faces are visible. There's asynchronous where you don't have a visible face, the dreaded telephone, which can actually be useful in certain circumstances. I guess you have to take all that into account.
Smith: I think that's true. I mentioned solution-ism earlier, where people are trying to find theapplication, or the platform that will solve all their problems. Deliberation is not a single thing. It's a bunch of different things happening; it's learning, generating ideas, listening, hearing and creating things together. I am suspicious of anyone who thinks that all of this can be done on one platform.
I think we might actually need to sequence the platforms. In face-to-face, we change the tasks that people do all the time and we change their relationships with each other and with the facilitators. In a way, it's like we're creating different platforms each time.
I wonder if we wouldn't need, for example, platforms that are specifically good at generating ideas and helping us visualise the argument space, and other platforms that are very good at allowing us to have some sort of face-to-face interaction, so that we can mimic some of the things we do on a small table. We might need another piece of software to start writing creative recommendations. We are able to manage this in a room by changing the way we use the space. I think we may have to change platforms, which again creates issues around the digital divide in terms of people's confidence to move from one platform to another.
Carson: I know that in NewDemocracy's deliberative processes, when participants write reports and develop recommendations, they usually use Google Docs and the group writes those documents themselves. We are very keen that the group has control over the resulting report, but this lends itself perfectly to an asynchronous environment. There's no reason why people can't all be working collaboratively on an online Google Doc at the same time.
Graham: That could be true. Although there are people who love being online and others who find it more of a chore. I worry about the "keyboard warriors". I think it's harder to deal with when you're not with people and you can't offer support to those who are perhaps a little more reluctant.
Part of this is what we do in mini deliberative audiences to support people who are less confident. We are able to see much more clearly how people interact with each other and support those who find it difficult. I'm not sure we can do that, when we only see a small picture of someone, and we only see their face.
In face-to-face situations, we see how people sit, how they move around the room, when they go for coffee, and whether people smile when they are not at the table. I think people who are not familiar with participatory processes may underestimate, for example, the importance of social time, the importance of looking at how people work, how they hold themselves. It's really hard to do online. There are all sorts of non-verbal actions that we observe. And this is also true for the participants, of course. They get signals, which you just can't get on Zoom or Skype or other platforms.
There is also a positive side to this. It can work well for people who are not particularly gregarious, social or outgoing, and who may be reluctant to talk in a face-to-face environment. They can be assertive online. As with any advantage, there is a disadvantage and vice versa.
Carson: On that point, people realise that when you are in front of a screen, you also have to take a break from time to time. I think we need to be very aware that screen time is not the same as face-to-face time. What was interesting with the French Climate Change Convention was that they recently spent a weekend online discussing the impact of COVID on climate change. And they had seven-hour days, as far as I can tell.
Smith: But it's interesting that people continued to do it. In the British Citizens' Assembly, the decision was not to do a full weekend. They're going to do three or four hour stints and put them together. We may well have to use the time differently online.
Carson: I think even four hours is a bit much. You talked earlier about the sequencing of events over a long period of time. We tend to do intensive sessions because there is usually a financial barrier to getting people together in a central location. There are advantages to this kind of intensive work, but I also think there are great advantages to doing it over a period of time and giving people time to think, to choose, to process and to do their own research.
Smith: Yes, I think that's right. My only concern - and this is an empirical question that we need to experiment and find out - is whether we're going to get the same volume of activity that we get with the types of mini-audiences that we know. The retention rate is generally amazing with these processes. But I think part of it is the social aspect, the fact that you are working with and meeting new people, building new relationships. I just wonder if they will be the same online. It's an empirical question. I don't know if they will be the same online.
One of the advantages of the French Citizens' Climate Convention and the British Climate Assembly is that they have done a lot of weekends before. So these people are already committed and have developed a collective work ethic. Can we build that kind of ethic online from the start, so that people feel that commitment in the process? I don't have an answer at the moment.
A lot of experimental work has been done online. There is a drop in participation, but that's usually the case with open processes where everyone can participate. We don't really know if you select a representative group and do the same kind of work that we would do with mini-audiences in person (telling them how important it is and explaining the kind of relationship it will have to decision making, letting them know that they've been selected and that it's a really special occasion), whether that's enough to keep them there or whether hanging out with people physically is really important.
Carson: Size is another issue to consider. We've worked with groups of 35 to 45 people in a similar process to a jury. And I think that's too many when we go online, that we might be better off with 25. And as you say, it's all experimental, we don't know until we do it.
Smith: It allows you to be much more experimental. We usually get everyone together at the same time. There's no reason why we can't get small groups online at different times that suit them. It's a matter of design, as we said before.
Carson: One facilitator told me that when he came to the end of an online session with people who were completely unfamiliar with the platform and needed a lot of support, he was surprised by the enthusiasm at the end. Participants said, "It was actually great for me, I learned something I would have been reluctant to learn.
Smith: Face to face, people say the same thing. At the beginning you hear, "Why am I here, I'm not going to be able to do anything. "And at the end, they have a high degree of political effectiveness.
One thing we haven't mentioned is that there are people who don't have much bandwidth; they don't have the technology. So part of the process is to provide that connectivity to people and teach them how to use the technology if they need to.
This was done in France and the UK. They found, for example, that some people had a computer at home, but it was being used by someone else for work. There are new barriers for us.
Carson: I think we've covered everything I wanted, but is there anything else I've missed?
Smith: One of the things I find quite exciting about online engagement is the use of argument viewing platforms. We haven't made the most of it face-to-face. We don't always map all the arguments that are out there. I think this can potentially lead to some arguments being overlooked, not deliberately. There is a real possibility of crowdsourcing - what are the arguments in this space? This is an online technology that we could use face to face. One of the interesting things that could happen here is that we do all these experiments online and then feed some of it into our face-to-face work.
I have a prejudice that face-to-face is better in terms of the deliberative process. I've always had this suspicion about online engagement. Some of this is due to the dysfunction of online spaces, but some of it, I have to be honest, is also just my own prejudice based on my familiarity with face-to-face processes. What I find really useful here is trying things out, experimenting with things and thinking, "This is actually really interesting. This works better than I expected". So, for me, it's going to be about that mix - how do you get the face-to-face and the internet to coexist more creatively?
Carson: It was so good to hear Graham Smith's perspective. He's right that the pandemic has forced very focused conversations between civil society technicians and deliberative designers in a very productive way. I like what he said about overcoming the digital divide, avoiding solution-ism, but also that there are exciting opportunities for experimentation.
This article is a translation by Open Source Politics of the article published on the Medium "Participo", an OECD publication. To read the original article by Lyn Carson, click here.
Recent Comments