Skip to navigation – Site map
2. AI in the City, the Age of Prediction and Anticipation

The Role of (augmented) Collective Intelligence for Municipal Governance

Interview of Frank Escoubes by Nicolas Miailhe & Arohi Jain
Frank Escoubes, Nicolas Miailhe and Arohi Jain
p. 47-49


In municipal governance around the world, the use of collective intelligence methods with dedicated tools and platforms is becoming the norm as a way to involve citizens, users and stakeholders in the design and implementation of policies. This new “open policy making” approach stands to benefit from the rise of artificial intelligence which can act as a cognitive agent to organize and summarize content, as well as a social agent interacting directly with participants. AI can also help fact check information and help generate automatic summaries and map concepts.

Top of page

Full text


1Assembl is the first deliberative online platform specifically designed to facilitate massive collective intelligence. Based on a multiphase consultation approach, it helps mobilize many people to tackle a complex issue. Assembl focuses on collective argumentation, dynamic structuring of ideas and noise reduction. It aims to co-create a strategic deliverable within a short period of time, usually between six to ten weeks. The tool and its supporting methodology are developed by bluenove. Assembl originated with a partnership with MIT as part of an R&D program funded by the European Commission. This open source software is widely applicable to large businesses as well as public entities and civic groups.

2The collective intelligence methodology enables creation of knowledge through interactions between community members and optimizes their engagement based on innovative facilitation roles. The motive is to structure the co-production of new knowledge. The platform allows for categorization, curation and synthesis of incoming messages within a formalized deliverable. This is conducted through a multi-stage process that is designed to promote deep content and dynamic structuring of ideas.

3Assembl works to reduce noise and focus community contributor’s attention to solve complex issues. This is done by organising and implementing four key facilitation roles within the platform; Harvester, the curator and extractor of ideas; Synthesizer, the creator of periodical synthesis on the proposals put forwards; Facilitator, the undertaker of community management, and; Knowledge Manager, the conductor of regular fact checking on the discussion content.

4The consultation is structured as a four-phase methodology that takes the community through stages of Sense-making, Ideation, Exploration and Prioritization. The multi-phase consultation is framed around a mind map of key thematic that lie within the scope of the debate. During the Sense-making stage, participants can contribute through a simple multilingual open question interface online. They are also able to view other contributor’s proposals and vote on these. Next, during the Ideation phase, the interface augments to a forum module enabling the community to have deep conversations on specific issues. This stage of the collective intelligence consultation is structured to engage the participants in a deliberative manner and develop well thought out insights. Third, the Exploration phase operates on a canvas module whereby a specific subject matter is structured in a polarised method to deepen the discussion. This acts as a funnel for the contributor’s opinion on a topic as the choice is binary. Lastly, in the Prioritization phase, participants are given voting tokens with which they can express their preference for certain propositions drawn out of the prior phases. Ultimately, the outcomes with the most tokens voted on results in a set of actionable proposals that are crowd sourced and collectively supported by the community.

5Given the rise of Artificial Intelligence methods in automatically analysing text, through Natural Language Processing (NLP) algorithms for example, Assembl is now increasingly relying on automated techniques to manage the gathering of collective intelligence. This is currently applicable with the functions of language translation as well as the role of the Harvester in collating discussion insights, and can be applied to many other aspects of the debate. AI can help progress collective intelligence through fact-checking information presented in the debate and ensure that participants remain well informed by intelligent crawling of data. In applying such artificial intelligence techniques, the process of mobilizing communities is augmented towards a more meaningful engagement at a larger scale, without the incurrence of substantial costs and time resources.

Nicolas Miailhe: What is collective intelligence and how can we use it with artificial intelligence technology to help revolutionise municipal governance?

Frank Escoubes: Collective intelligence is the ability to mobilize large communities of people to co-design solutions to complex issues. Co-creation means combination and multiplication of perspectives. Such methodologies require both an iterative process of new knowledge design and massive scale: tens or hundreds of thousands of people providing rich insights that inform and orientate public policy.

In municipal governance, the use of collective intelligence is critical to understanding how all interested parties can help co-design policy recommendations. This can also be referred to as “open policy making”, a notion closely related to the idea of deliberative democracy (inspired by Habermas among others). I believe that this is the only effective way to convene citizens as experts of their own contextual lives, and therefore as legitimate providers of inputs that are required to imagine the policies and programs of tomorrow.

Collective intelligence is indeed best described as deliberative democracy, where quality of ideas is somehow the end goal, whereas participatory democracy hinges upon the quantity of participants, most often failing to gather profound and heterodox thinking. I believe deliberative democracy will be the supporting paradigm behind future municipal governance. Of course, the quest for scale is calling for supporting AI-enhanced methodologies.

Arohi Jain: What are the key opportunities and benefits of using collective intelligence to guide governments on public opinion?

Frank: Many people tend to think of collective intelligence applied to democracy in the restrictive context of the legislative process. I, on the other hand, consider it to be highly relevant and applicable to the entire spectrum of public policy making, covering laws, policies and programs. It should also cover all co-design stages of policy making, from evidence-based diagnosis to collective ideation to policy recommendations to policy evaluation. Each stage requires varied levels of maturity of contributors. It is also very important to not only mobilize citizens, who express themselves in their own names, but also all constituencies in civil society (non-profit organisations, NGOs, pressure groups, industry organizations, etc.) who represent the consolidated point of views of a given set of stakeholders.

Nicolas: In your experience, what have been the challenges in using this methodology to inform public policy?

Frank: The greatest challenge of collective intelligence lies in engaging citizen. It is extremely difficult to mobilize a large range of the population that has a diversified base of knowledge on specific public policy issue. This is compounded by the challenge of finding the right balance of key stakeholders and citizens to provide political legitimacy to the consultation.

Secondly, the digital divide factor is still a reality. Online consultations that are used to garner insights do have significant entry barriers. They need to be fertilized with offline events (interviews, workshops, meetings, World Cafes, beta-tests, etc.). This has operational and logistical consequences: open democracy is to be considered hybrid from the get go.

The final challenge is of a cognitive nature. In assembling the full spectrum of participants for a civic consultation, the difficulty lies in ensuring whether the stakeholders have access to the right level of information for a productive consultation. This could further imply that an educational exercise is needed prior to the consultation. The role of experts should not be underestimated as well. Open democracy is first and foremost a citizen training process and it has to be reconciled with the world of experts for insights, data, evidence-rich analyses, complex interpretation, scenario-planning, etc.

Arohi: How will the rising trends and drivers of artificial intelligence impact the way we gather collective intelligence?

Frank: There are several ways to employ artificial intelligence depending on the context of the consultation. In our case, we use a deliberative platform, Assembl, structured around threads of discussion that organically grow in and around various themes; the challenge lies therefore with natural language processing and generation. Over and above certain thresholds of participants, the multiplicity and diversity of user generated content calls for narrative text analysis through machine learning algorithms. In this case, artificial intelligence poses significant benefit in acting as a cognitive agent that can organize and summarize content (a.k.a. knowledge creation systems), as well as a social agent interacting directly with participants, through chatbots and virtual assistants, hence community activities.

Furthermore, artificial intelligence can help fact checking information presented in a collective intelligence exercise. In dealing with large number of people on a specific consultation, AI can ensure the participants remain well informed through intelligent crawling of data libraries and enable support or challenge the views automatically. This capability would otherwise be extremely time consuming for community managers.

Lastly, with the detail and amount of content generated while participants are discussing a specific topic during a debate, it can be challenging to keep everyone up to date. Here artificial intelligence can provide an elegant solution by generating automatic summaries of the debate and mapping concepts that provide participants with easily accessible capsule updates on the discussion.

Nicolas: How do you see the use of AI in collective intelligence evolving over the next three to five and then ten years?

Frank: Well, this is a difficult question given the challenge in understanding the evolution of AI itself! I believe there is great potential in using AI to reconcile public policy making by citizens and data analysis. In the medium term, we could design a data-centric collective intelligence system that uses the power of data interpretation by algorithms to nurture, inspire and navigate creative human recommendations. I suspect this will happen in the next 10 years.

Another element is related to how we use AI to enhance creativity of citizens. It is currently difficult to radically shift to new societal mechanisms and therefore if we could create a way in which AI could feed efficiently the creativity and co-designing processing for citizens that would be exciting.

Top of page


Bibliographical reference

Frank Escoubes, Nicolas Miailhe and Arohi Jain, « The Role of (augmented) Collective Intelligence for Municipal Governance », Field Actions Science Reports, Special Issue 17 | 2017, 47-49.

Electronic reference

Frank Escoubes, Nicolas Miailhe and Arohi Jain, « The Role of (augmented) Collective Intelligence for Municipal Governance », Field Actions Science Reports [Online], Special Issue 17 | 2017, Online since 31 December 2017, connection on 20 November 2019. URL :

Top of page

About the authors

Frank Escoubes

Frank Escoubes is the Co-President and Founder of Bluenove. He has 20 years of experience in strategy consulting, including 8 years at Deloitte in Canada. In 2011, Frank created the collaborative web platform Imagination for People dedicated to social and societal innovation, of which Bluenove became the key financial partner. Frank joined bluenove in 2014 as Executive Chairman. He is particularly responsible for developments in collective intelligence through the launch of Assembl software and partnerships with major international institutions (European Commission, MIT, etc.). Passionate about creativity and economic development, Frank accompanies companies, clusters and cities around the world in the reinvention of their strategic framework. He actively puts in place the dynamics of collective intelligence and sources world experts. Frank has been an Ashoka Fellow since 2012.

Nicolas Miailhe

By this author

Arohi Jain

Top of page


Creative Commons Attribution 3.0 License

Top of page
  • OpenEdition Journals