Can Collective Intelligence Enhance Democracy?

Research Proposal - MSc in Digital Politics and Sustainable Development

*Please note, this account has shifted to Breathe with Bella content - for more content like the below, please subscribe to our sister account The ANTIPARTY Journal:

Section 1 - Introduction

Democracy is fundamental to a well-functioning society. The term refers very generally to a method of collective decision-making, its etymological origins rooted in the ancient Greek dēmos ‘the people’ and kratia ‘rule’, quite literally translated by Abraham Lincoln when he called for “government of the people, by the people, for the people”. Yet across the globe, democracy is under threat – the International Institute for Democracy and Electoral Assistance (IDEA) states that more than a quarter of the world’s population currently live in countries where democracy is in decline and the Economist Intelligence Unit found substantially more than a third of the world’s population live under authoritarian rule, a trend worsened by the pandemic. The pro-democracy organisation Freedom House pointed to 15 years straight of steady democratic backsliding, warning that the scales of global freedom are now tipping in favour of authoritarian-leaning populists and would-be dictators. Indeed, IDEA found for the fifth consecutive year, the number of countries moving towards authoritarianism is approximately three times as high as the number moving towards democracy.

Protecting our democracy is of huge significance now more than ever and in December 2021, US President Joe Biden held a two-day Summit for Democracy with 80 world leaders, seeking to address the issue. However, he failed to invite Russia and China, whose respective ambassadors to the United States issued a joint statement noting that “democracy is not just about domestic governance; it should also be reflected in international relations”. A few mere months after the event, Russia invaded Ukraine and the conflict is ongoing at the time of writing. Thus, maintaining a healthy democracy is not just important towards societal well-being but it is also critical on the geopolitical sphere with implications on human security.

For the purposes of this paper, focus will be placed on the instrumental values of democracy – it is instrumental by way of encouraging people to be more autonomous via political decision-making in order to elect representatives that will best reflect our needs in resultant laws and policies. As we shall see later, this sense of autonomy links closely with having agency and a level of control. Whilst the UK maintains a relatively high ranking on international measurements of democracy, polling of UK citizens consistently returns high levels of dissatisfaction with the democratic system. Increasingly this dissatisfaction centres around the honesty and integrity of elected representatives, which feeds into overall levels of trust. If we cannot trust those who govern us to do a good job, the people should be able to remove them, as noted by the Founding Fathers of the US Constitution who stated in 1776 that “when any government shall be found inadequate… a majority of the community hath an… indefeasible right to reform, alter, or abolish it.” However at the time of writing the sitting Prime Minister is the first in history to have committed a criminal offence and one of his ministers has been charged with sexually assaulting a minor, yet he retains his post as a public official. These non-exhaustive, but significant examples threaten our democracy as a system of government which should allow the selection and replacement of the government through free and fair elections.

In January 2022, The Constitution Unit of University College London (UCL) published the results of a major survey of public opinion fielded from a sample of 6,500 people which were deemed to be a representative of the voting age population across the whole UK.¹ When asked what they viewed to be the most important feature in a democracy, the public confirmed this to be the ability to vote out those in power who do a poor job (pg. 8, Renwick et. al, 2022).² The majority of respondents also viewed democracy as contingently good ‘so long as it delivers effective government.’

However, when asked how much they trusted the four central parts of the democratic system, the Prime Minister received the most scores for ‘strongly distrust’ (pg. 2, Renwick et. al, 2022) meanwhile ‘being honest’ was ranked the most important characteristic for a politician (pg.4) above ‘getting things done’. This is echoed later in the survey where integrity is chosen over delivery with regards to policy that is best for the country and policy that most people want (pg.5). This demonstrates the crisis of trust felt by the UK population where simply acting with integrity³ is preferred over delivering on outcomes. For want of democratic tools to be able to exercise the instrumental value of democracy when those in power are no longer deemed fit to run the country, an alternative solution emerges which shifts the focus from a failing representative democracy to one that is more direct and deliberative.

Indeed, over three quarters of the UK population thought they had too little influence on how the UK is governed (pg. 12, Renwick et. al, 2022) and wants the system to be more responsive to citizens such as MPs following their constituents’ wishes. Crucially, the majority of respondents, across most demographic groups and past vote choices, supported the use of Citizen’s Assemblies, defined as “a group of people are selected at random, in much the same way as for jury service. Organisers try to make sure people of different ages, genders, ethnicities, class backgrounds and political views are represented. These people are given information about the issue and hear different arguments. They get to ask questions, think about the evidence, and discuss different views among themselves. Then they vote on what they think and their conclusions are made public” (pg.13). This links to the understanding of democracy to be a method of collective decision-making.

It has been demonstrated that the UK public are willing to move towards a more involved, participatory democracy but interestingly it is also shown that most people believe it is both equally important to follow the will of the majority and protect the rights of minorities. As we shall see in the next section, this leaning towards consensus and finding middle ground over an absolute majority carves room for the field of collective intelligence.

Fig. 1 –Survey responses from the Constitution Unit report What Kind of Democracy Do People Want?

This paper will focus on democracy within the UK as it provides the greatest opportunity for democratic innovation alongside high levels of transparency and access to data. Further, mass-level liberty aspirations are shown to have a positive impact on democratic change – even stronger than economic modernisation (Welzel, C., & Inglehart, R., 2001) suggesting that where there is a will, there is a way. The nascent field of collective intelligence is thus proposed as one way to enhance the instrumental values of our democracy and this paper sets out a method of generating this through a form of citizen’s assembly, enabled by technology.

Section 2 – Literature Review

Deliberative Democracy

Deliberative democracy focuses on decision-making by deliberation, distinct from traditional democratic theory which relies on voting. Its principles have been built upon by the works of German philosopher Jürgen Habermas on communicative rationality in the ‘public sphere’, a space for citizens to reach consensus free from distortions of unequal political power (Habermas: 1974; 1994; 1998; 2005; 2006), thus upholding the intrinsic democratic value of equality.

Political theorists Bruce Ackerman and James Fishkin set the scene for deliberative democracy in action through their ambitious proposal for a ‘Deliberation Day’, intended to take place during a public holiday two weeks before each general election which hoped to dramatically improve citizen participation. Whilst such large-scale deliberative interventions have to date remained practically unachievable, the underlying principles for such ambitious proposals has manifested in the growth of Citizen’s Assemblies in recent years. Fishkin’s definition of deliberative democracy includes citizen participation in the decision-making process, describing five characteristics essential for legitimate deliberation thus operating as part of a direct democracy:

• Information: The extent to which participants are given access to reasonably accurate information that they believe to be relevant to the issue

• Substantive balance: The extent to which arguments offered by one side or from one perspective are answered by considerations offered by those who hold other perspectives

• Diversity: The extent to which the major position in the public are represented by participants in the discussion

• Conscientiousness: The extent to which participants sincerely weigh the merits of the arguments

• Equal consideration: The extent to which arguments offered by all participants are considered on the merits regardless of which participants offer them

Citizen’s Assemblies have become increasingly popular towards debate around the environment as the climate crisis has held top positions on the global agenda. As well as being a topic which people are passionate about, it is noted to be important for scientific choices that involve both ethics and highly specialised knowledge be deliberated in a public forum so to educate the public and legitimise decisions (pg. 189, Mulgan 2018). However, drawing on extensive American survey data, Yale professor of psychology Dan Kahan and his colleagues reveal that public concerns about climate do not correlate with scientific knowledge or reasoning capacity. There runs the risk of emotional rather than rational deliberation feeding into legislation so this study will continue to assume a representative democracy and take into account the preference of the UK public to not cut out elected representatives completely.

Fishkin’s five characteristics above will be factored into the design of the research methodology as well as the considerations of more recent UK-based practical examples of deliberative democratic participation in practice (Pearse, 2020; 2021).

Collective Intelligence

Collective intelligence, as defined by the Centre for Collective Intelligence Design, is created when people work together often with the help of technology, to mobilise a wider range of information, ideas and insights to address a social challenge. The solution formed by consensus-building by participants has been proven to be more satisfactory than traditional voting methods (Grobbink & Peach, 2020) and produce greater impact (LocalGov, 2022). It can therefore be seen to strike at the heart of what it means to have a democracy, defined as a form of collective decision-making and more specifically a deliberative democracy where citizens form a collective with a common good that serves the basis for public deliberation (Ball, 2021). It taps into the wisdom of crowds, requiring a high degree of diversity and can flexibly adapt to the challenges of our complex world (Helbing, D et. al, 2018), now made possible in practice through the use of artificial intelligence.

Intelligence, on the other hand has many definitions but its etymological roots derive from the Latin inter, meaning “between” combined with the word legere, meaning to “choose”. It is therefore inextricably tied with making decisions and choosing between paths, linking to the concept of freedom of choice. Freedom is a core part of democracy’s intrinsic values and is therefore something to aspire towards and protect. Intelligence as we know is becoming increasingly artificial and as this artificial intelligence develops especially in the field of neural networks that seeks to recreate the processing of data akin to the human brain, it is imperative that we are able to retain agency not least in the functioning of our societal system. Thus, as artificial intelligence becomes increasingly integral to our everyday lives through our reliance on technology, data capture and storing of our preferences, we can retain agency in the functioning of our democracy by consciously participating in it.

Research into collective intelligence towards democratic means is to date few and far between but nascent, as it makes finding appropriate solutions to the most complex challenges of our world possible (Helbing, D et. al, 2018) such as humanity’s grandest challenges of climate, health or war, which we have little prospect in solving without progress in how we think and act together (Mulgan, 2018). Whilst collective decision-making as a concept is not new, it is the technological means at our disposal and the worsening political (as well as environmental) climate which makes the area well worth revisiting. Such enabling technology also serves to challenge those who hold the view that democracy is inefficient at best (Caplan, 2008; Brennan, 2017). Whilst investment into large-scale intelligence naturally flows to highly competitive fields such as finance and the military⁴, fields which offer much greater gains to society tends to miss out. Therefore, this project attempts to reverse this trend by channelling technological development towards democratic and civic innovation.

Collective intelligence stems from the premise that intelligence is distributed⁵ – not too dissimilar to the original ideas behind the internet which suggested we see distributed brains as assemblies, echoing Tim Berners-Lee’s description of the theory behind the worldwide web as made up of “abstract social machines” with “processes in which the people do the creative work and the machine does the administration” (Berners-Lee, 2000) which could describe the creative deliberation of a citizen’s assembly, the ideas from which are then passed on to parliament for legislating. This network of intelligence can also reflect distribution of power, which the UK population is consistently unwilling to concentrate in the hands of a few (pg. 15, Renwick et. al, 2022). This project seeks not to completely replace humans by machine capabilities in the functioning and participation of democratic process, but rather enhance human participation and deliberation via use of technology towards consensus-building for decision-making.

Section 3 - Research Methods

It has been demonstrated why collective intelligence offers a truly democratic, technologically enabled tool for decision-making by consensus. We have seen sparse but promising efforts in the space so far and this project intends to contribute towards creating a more unified field for collective intelligence in local governance and participation, by offering the foundations of a framework that can be adopted and evolved over time. This is not straightforward, as it requires active participation, deliberation and moderation as this next section will outline. However, it is for this very reason that efforts need to be directed towards technological solutions for augmenting, rather than automating decision-making in order for people to retain agency within our democracy. This is important also because the alternative of inaction, knowing the capabilities of social media algorithms and government behavioural ‘nudge’ units, could lead to unconscious control by malicious actors.

Before outlining the specific research methods to be adopted in this project it is worth reviewing a few examples of others with a similar aim. International case studies of digital deliberative democratic models which have been successful include the Your Priorities platform developed in Reykjavík, Iceland in 2016 which involved citizens proposing and voting on ideas to spend a multimillion-dollar budget. The site requires participants to write arguments for and against options, which the site then ranks – this speaks to the 65% of the UK public who believe that for big political questions, it is generally best to weigh the pros and cons of different options before deciding (pg.10 Renwick et. al, 2022). Other digital forms of deliberation in participatory budgeting have also been used in Madrid and Brazil (Carr, H et. al, 2020). Meanwhile, the East London borough of Tower Hamlets is currently consulting on survey responses it collected from residents as to how to spend this year’s budget,⁶ in what appears to be a laborious and opaque process.⁷

Swarm AI

Beyond budgetary considerations, digital democratic deliberation can take the form of ranking qualitative solutions in order of preference. Swarm AI, developed by Unanimous AI as part of Nesta’s Collective Intelligence grants programme found that algorithms modelled on the swarm behaviour of honeybees enabled politically polarised British voters to set government priorities in a more satisfactory manner than those produced by a traditional majority vote (Nesta, 2020). A total of 237 participants were recruited to rank lists of government priorities individually through a survey as well as together on the Swarm AI platform, which allows networked human groups to deliberate in real time. Distributed artificial intelligence worked together with human groups to form a closed-loop system where both the machine and individuals can react – algorithms were trained on the behavioural dynamics of groups and in the second step, a neural network model used the interaction dynamics to generate a conviction index. The index estimated the group’s confidence in the final outcome (Berditchevskaia & Baeck, 2020).

Fig.2 – The graphical puck within Unanimous AI’s platform Swarm AI

Whilst successful overall, the experiment found similarly satisfactory results between swarming and Borda count, a method of converting the rankings into a score, particularly on government issues rather than objectives and specifically on the issue of immigration. This supports the caution posed by Mulgan (pg. 188, 2020) in that issues involving deeply held beliefs may not be so conducive to thoughtful deliberation and online public airing can simply “raise the temperature”, as enabled by the real time functionality of Swarm AI’s platform. Hence there is the challenge of working out which kinds of issues are best suited to what kinds of engagement.

Pol.is

Casting the net more broadly, the vTaiwan project which sprung out of the Sunflower Revolution in 2014 allowed citizens of Taiwan to co-create legislation in a project that engaged hundreds of thousands of participants. The platform’s first test was to regulate Uber’s operations in the country and it did this through Pol.is, an open-source technology that leverages data science, enabling a real-time system for gathering, analysing and understanding what large groups of people think in their own words, through advanced statistics and machine learning. Some describe the platform as a social media network however this is misleading as conversely to major social media platforms, Pol.is does not include a reply button so participants can avoid trolling, trading insults or going down a rabbit hole. Instead, people compete to find the most nuanced statements that garners the most popularity so the initially disparate groups of opinion is seen to converge.

Pol.is gamifies finding consensus so participants spend their time discovering commonalities instead of outright majorities and after about a month debating the issue of Uber, four groups of opinion or ‘clusters’ would become two as items emerged that enjoyed near-unanimous support. One statement received 95% support across all groups, which helped to direct government action and since the issue of Uber, Pol.is has been used to set the agenda for numerous pieces of law and regulation on topics such as online alcohol sales and revenge pornography.

Fig.3 – The converging of clusters of opinion on the Pol.is platform

The limitations of Taiwan’s use of Pol.is includes its current application being limited to digital issues only and it has not yet been tested on issues of ‘real’ national division. To overcome this, the British cross-party think tank Demos recently piloted the use of Pol.is as a tool to facilitate democratic debate in the UK, where it explored public attitudes to data-driven political campaigning. The experiment ran for one week and it was found to have succeeded in surfacing areas of consensus and encouraging debate (Smith et. al, 2020). It was also the first experiment to conduct Pol.is using a nationally representative sample as opposed to a self-selecting sample, which included marginalised groups. This ensures that resulting policies will benefit from the ideas and experiences of groups who would otherwise be unlikely to take part in the process and make the findings more persuasive to policymakers and politicians (pg. 27, Smith et. al, 2020). However, the project noted the key difficulty to be the translation of outcomes generated by the platform into becoming policy and legislation. As a think tank, it is natural for Demos to raise this as an issue however it contrasts to the Constitution Unit’s survey where the public clearly did not want to cut out elected representatives from decision-making – they found support was highest for a citizen’s assembly whose recommendations would be passed on to parliament,⁸ and lowest for an assembly whose proposals would automatically become law (pg. 13, Renwick et. al, 2022).

Demos also notes this to be a key difficulty with “all” democratic innovations, yet as demonstrated by the use of Pol.is in Taiwan since 2014, this is not a universal difficulty but instead reflects a key difference in the support granted by the national government in question. As Colin Megill, one of the founders of Pol.is observed, "change the information structure, and you can tweak power". Perhaps this is why the UK government has not been as inviting to democratic innovators as in Taiwan, however lest we forget the latter was catalysed by a revolution.

Methodology

The field of collective intelligence, as noted by Mulgan (pg. 4, 2018) “needs to be both open and empirical. Just as cognitive science has drawn on many sources… to understand how people think, so will a new discipline concerned with thought on larger scales need to draw on many disciplines, from social psychology to computer science.” This research draws on the intersectionality between the fields of digital politics, responsible artificial intelligence and sustainable development, under the umbrella of social science. It will adopt a hybrid approach of collecting both qualitative data in the form of opinions and preferences as well as quantitative data in translating these sentiments into scores. Within the research methods there will be an element of computer science from a purely practical perspective of analysing the data gathered. Given the success of the Pol.is platform and its availability, this offers the most likely tool to be used to gather primary data however as the software is open-source, this research remains flexible to replicating a similar, more tailored model that also processes data more securely.

When dealing with a social science, the scientific method deployed must be robust in order to test the hypothesis of the research in the most objective manner as is possible. To this end, the researcher will be made to reflect on any potential biases and ensure that one’s own interests it not an obstacle to such controlled enquiry. In the same vein, any conflicts of interest found with the subject(s) of the research will also be declared. In light of Weber’s answer to the positivist thesis, as the research relates to a normative question the researcher will also minimise the impact of one’s own values on the research by being aware of and declaring any influencing factors at every stage. For example, the desire for a certain outcome will not modify the observation stage of the research towards one’s own expectations. It is accepted that this research is but a feasibility study, not a final solution and nor is it the only tool in isolation towards the overarching aim of enhancing democracy.

Based on related experiments reviewed so far in the field, it is proposed that the chosen subject area for deliberation involves a proposition rather than an issue in order to contribute towards government agenda-setting. It should not involve deeply held beliefs, to allow fluidity for consensus-building which can be gamified. It follows that a fruitful exercise would be to construct a tool for participatory budgeting similar to those shown to work in Iceland, Madrid and Brazil (Carr et.al, 2020); one which also responds to the observed willingness of local councils such as Tower Hamlets to involve the population in this process. This research can offer a platform where citizens’ preferences are captured in a more secure, efficient and overall democratic way that saves the local council valuable time spent consulting on manually collected responses from traditional methods. The benefits of this approach is two-fold as participants are likely to have enhanced levels of satisfaction due to their proximity to potential outcomes felt in their local community, further to their pre-existing investment in the exercise by virtue of paying council tax.⁹ As noted by Pearse (2020), “deliberative democratic exercises… result in judgements that are legitimate if arrived at through discussion involving the people due to be affected by the resultant policy.”

Testing the Hypothesis

A decision arrived at by consensus generated through collective intelligence in a deliberative democratic facility such as a citizen’s assembly improves the instrumental value of democracy measured by levels of satisfaction, secondly by legislative adoption. A technologically-enabled forum produces satisfactory consensus quicker but absolute levels of satisfaction can be lower than levels of satisfaction achieved in-person.

A form of the Pol.is open code source data will be used in addition to survey responses which will be used to formulate the initial set of statements as well as collect feedback as to levels of satisfaction at the end of the experiment. Should technological capabilities allow, this study is open to exploring real-time capture of data to overcome issues of representation via access to digital tools, thus the research is open to facilitating a live discussion.

This tests the political participation operational dimension of democracy and touches on re-imagining deliberative democracy towards an epistemic democracy, which has been coined to capture this sense of democracy as a way of generating knowledge and tapping the wisdom of citizens, not just choosing between alternatives (pg. 183, Mulgan 2020).

Whilst the main indicator is satisfaction by consensus, the next stage would be to follow this through to implementation which continues to be the responsibility of elected representatives in a representative rather than direct democracy. Re-imagining a system around existing institutions only goes so far. However, there is the possibility that the research may also contribute to forms of organisational decision-making where the firm wishes to include more democratic measures and qualitative feedback to its strategic planning – although there is less structural impetus to do so in a typical for-profit, there is scope to contribute to public sector and civil service organisations.

Potential Limitations

There are various limitations to be considered in carrying out a research project of this size. Firstly, due to the lack of financial incentive offered to volunteers, there may be issues in the acquisition and retention of willing participants throughout the duration of the study – there can be high rates of attrition during the process (French & Laver: 2009; Kenyon: 2005). Therefore, a subjects’ ongoing engagement depending solely on goodwill may result in a bias towards people with particularly strong views on the subject being deliberated, or those who have a preference to make the experiment a success. This strengthens the proposal to focus the experiment on a local council’s budgetary spend to which the subject is already contributing. Following on from this, it is recognised that the initial trial will not be a fully representative sample of the population.

Another caution would be to ensure that the study does not recreate existing power imbalances which exist in society – involving a specialised subject for deliberation risks an expert-centric process that legitimises existing policy hegemonies and prevailing wisdoms, which would go against the theoretical advantage of collective intelligence being that groups of diverse people are collectively smarter than any single individual.

Ultimately, as noted in prior studies in the area of democratic innovation, it must be factored into the research that any output may not directly translate to legislative or regulatory outcomes. Thus, this proposal can seek to provide a meaningful method of enhancing democracy via increased civic participation and satisfaction of the populate. At best, it will form a basis to scale a similar methodology and process for digitally capturing and formulating consensus amongst a representative sample of a local community. In doing so, it responds to calls such as that by Kathy Peach, Co-Director of the Centre of Collective Intelligence Design for solutions to “work with institutions, work with communities and use technology to enable those positive collective behaviours”.

Ethical Considerations

When harnessing the use of technology in participatory projects of any scale, it is important to consider the ethical implications in order for the application of collective intelligence to thrive in the area of citizen science. This project seeks to take as much ethical considerations into account as is relevant in order to amplify the potential solutions achieved towards a participatory digital democracy whilst also shaping the future of responsible technology and machine intelligence to support society.

The research carried out will conform to the basic principles of beneficence and non-malfeasance. All participants in the study will be adequately informed of the aims, methods, benefits and hazards potentially arising from participation. A written copy of their informed consent will be collected and recorded, which will include their freedom to withdraw from the study at any time as well as confirming their status as volunteers, so that any decisions not to participate does not prejudice the subject in any way.

Data Privacy

This study will seek to ensure that all data is obtained and processed fairly and lawfully. Data should be adequate, relevant and not excessive for the purposes of this research nor will it be kept for longer than is necessary. Where necessary, wider ethical codes such as the UKRI Research Ethics Guideline will be observed. Considering a recent decision by the European Court of Justice which means the use of US hosted servers for EU citizens’ data in the future could be unlawful (Schrems II: 2020), the open-source code of Pol.is is understood to mitigate against the co-option of the technology by private interests. This is perhaps the key benefit to such a method; not only does it enable a democratic event, but the innovation democratises the process of conducting the event itself (Smith et. al, 2020).

“What is possible in collective intelligence might advance over the course of the next century, given advancements in machine learning, and how we can enable those who seek to empower citizens to operationalize those advancements in policy-making processes. Ultimately, we seek a world in which political agenda-setting power rests in the hands of the most inclusive possible set of engaged citizens and independent public servants, rather than in the hands of the concentrated few. Democracy needs to evolve to survive and as our lives become increasingly digital, so does our methods of deliberation and participation in a well-functioning democracy.”

1. The project is part of its ‘Governance after Brexit’ research programme and includes a Citizen’s Assembly on Democracy in the UK and a follow-up survey of the UK population to be conducted in spring 2022. Further reports are expected later in the year.

2. Interestingly, being active in their local communities was seen to be least important, especially in direct comparison to the aforementioned – we will revisit this later in the research methodology.

3. Acting with integrity should be a given for public office holders based on The Seven Principles of Public Life, also known as the Nolan Principles.

4. For example, Primer AI is a private company whose client base includes the US Airforce and US Army. It uses Natural Language Processing (NLP) and enables early warning systems as a means to constantly surveil the information manipulation landscape towards the threat of disinformation as a national security concern. At the time of writing, Primer AI’s total funding amount is $160 million.

5. Beyond the scope of this project but it is worth acknowledging the trend towards distributed ledgers of the blockchain being increasingly used for transactions and the current shift towards the purportedly decentralised ‘Web3’ where people own their data, not corporations. This project would suit such an environment and is therefore open to eventually integrating with Web3 in moving towards a digital democracy.

6. https://talk.towerhamlets.gov.uk/budget

7. Attempts to register for updates on the current proposals or view the results of previous years’ consultations returns an error 404 ‘page not found’ online.

8. Similar to a petition receiving over 100,000 signatures, passing to parliament does not mean that the issue will be taken seriously as opposed to an act of procedure. In Iceland, its national parliament involved the public in an open process to rewrite the constitution, with online inputs and a representative commission. The rewritten constitution was then endorsed in a referendum, yet ultimately rejected by the Parliament after a general election had changed its makeup (pg. 186; Mulgan 2018). Thus without any enshrined equivalent for parliament to consider the outputs of a Citizen’s Assembly, there runs the risk of wasted effort.

9. In London as announced in December 2021, the GLA ‘precept’ will increase by a record 8.8% from April 2022. Providing local residents an avenue to express their preferences on how their increased council tax payments will be spent could also counter-balance any issues with volunteer acquisition as discussed below.