1. What Are the Debates in the Deliberative Democracy Community in Relation to Technology?

Technology has recently entered the scene of deliberative democracy, both as a subject of deliberation and as a way to improve deliberative processes. In the recent Artificial Intelligence (AI) Action Summit held in Paris in February 2025, Missions Publiques (2025) and the Stanford Deliberative Democracy Lab launched the Global Coalition for Inclusive AI Governance. This initiative aims at bringing together more than 10,000 citizens from across the world to deliberate about the future of AI governance. In Flanders (Belgium), the ‘amai!’ programme uses civic engagement tools such as crowdsourcing and citizen juries to design and fund narrow-specific AI projects, to broaden the audiences involved in the development of new technologies (Duerinckx et al. 2024). Big Tech platforms are increasingly organising their own deliberative projects (Malkin & Alnemr 2024). In the European Union (EU) context, a European Citizens’ Panel (ECP) on the topic of ‘Virtual Worlds’ was held in 2023, followed by another on AI in 2024.

In parallel, the growing momentum of deliberative minipublics on the regulation and governance of technology is matched by the increasing use of technology as a way to improve citizen assemblies (CAs). There is increasing literature by both academics and practitioners about the possibilities of using technology to improve deliberation. The mainstream position in this debate appears to be that, while it is not a bulletproof solution, technology holds the potential to make deliberative democracy better, by integrating AI into citizens’ assemblies (McKinney 2024) or by using it to bring deliberation to the masses (Landemore 2024). What these innovations have in common is a belief in technological solutionism, a concept advanced by Evgeny Morozov (2013), which suggests that societal problems can be ‘solved’ through technology. In the specific case of deliberative democracy, it is the belief that technology is an ‘efficient’ way of solving the ‘problems’ of deliberative minipublics. Indeed, Goñi (2025) has explored how, while the deliberative democracy literature carries on such ‘technological solutionism’, the science and technology studies (STS) literature engages in a sort of ‘participatory solutionism’ in the design of new technological artefacts. Relatedly, deliberative minipublics are seen as a ‘new paradigm of democracy’ (Landemore 2020) that can better connect citizens with political elites and consequently further democratise democracy. However, the emergence of this so-called ‘deliberative wave’ is also the result of lobbying efforts by deliberative democracy consultants.

First, we take stock of the latest developments in CAs, focusing specifically on the ECPs and Google’s Habermas Machine (HM). We argue that introducing technology as a ‘solution’ to ‘fix’ some of the ‘problems’ within the deliberative democracy community reinforces the depoliticisation and disintermediation of deliberative initiatives. Therefore, the focus on the technological component distracts from criticising the problematic nature of minipublics from a systemic understanding of deliberation.

2. The European citizens’ panels and the Habermas Machine

3. If Technosolutionism Harms Democracy and Deliberation, What Are the Alternatives?

While in STS literature, citizen participation is often seen as the ‘solution’ to the ‘problem’ of technology, in deliberative democracy theory, technological innovations arise as the ‘solution’ to the ‘problem’ of citizen participation (Goñi 2025). Technosolutionism applied to deliberation assumes that the problem with minipublics is that their nature hinders their scalability (McKinney & Chwalisz 2025) or other ‘solvable’ technical problems, ignoring structural limitations. Although AI can multiply participants, streamline processes and reduce organisational costs, the use of AI does not solve the two main problems of minipublics: the inherent depoliticisation and individualisation of citizen participation (Oleart 2023b). The technosolutionist perspective adopted by many practitioners of deliberative minipublics and scholars naturalises the relationship between deliberative democracy and minipublics, while sidelining the role of conflict and collective actors. By conceiving of the ‘problems’ in small and technical ways (scalability, sortition, moderation, translation, inclusion, speed, etc.), it misses the wood for the trees. It moves the discussion away from the debate on the systemic desirability of minipublics, sidelines a mass politics conception of democracy, and leaves unquestioned the political economy of AI, owned overwhelmingly by a small group of Big Tech companies. Furthermore, it contributes to the disintermediation of participation, as participation becomes increasingly mediated by technology. With AI technosolutionism, it is possible to participate as individuals in ‘deliberative’ processes without actually engaging with collective actors or in collective discussions. AI may become the new mediator that channels the voice of ‘everyday people’, reinforcing the commodification of minipublics where deliberation is turned into a product to be sold. Indeed, Big Tech platforms appear to be interested in developing deliberative ‘products’ such as the HM as a way to ‘citizenwash’ their corporate profit-seeking nature, and deliberative consultancies are likely to integrate AI because it facilitates their work. The process of marketisation continues its cycle: consultants sell CAs to governments and institutions, and Big Tech entrenches itself in yet another area of society.

Democratising the relationship between technology and participation entails reclaiming a systemic understanding of deliberation, rather than focusing on the specific procedures or technical methods that ‘improve’ deliberation within minipublics. Therefore, the way to materialise deliberation at the systemic level is not, as Landemore (2024) proposes, scaling up minipublics to cover an important but insufficient part of the population. The focus ought to be oriented on embedding deliberative initiatives to foster an agonistic public sphere (Mouffe 2013), connecting with relevant mediators such as political parties, civil society, trade unions or social movements. This also means questioning some of the premises of deliberative minipublics. Rather than aiming for the ‘descriptive representation’ of sortition, perhaps we should be thinking about the kind of institutions that can channel the existing structural conflicts that exist in society to reverse material inequalities. A more meaningful problematisation asks: how can we empower the social and political groups that are structurally absent from the public sphere? Relatedly, considering the increasing business of deliberative consultancies and the private ownership of AI, how can we democratise the political economy upon which deliberative initiatives are built?

From this perspective, technological ‘solutions’ to ‘improve’ deliberative democracy appear as primarily yet another ‘shortcut’ (Lafont 2020) that moves attention away from articulating a real democratic infrastructure based on vibrant debates in the public sphere and strong collective actors. It is certainly tempting to imagine that democracy has technological ‘fixes’ or ‘solutions’ in the context of the void left by hollowing out of mass membership organisations (Mair 2013), especially in the EU context, where there is a longstanding consensus-oriented understanding of deliberation and democracy (Crespy 2014; Oleart & Theuns 2023). However, such temptations are misguided. Not only does the HM perpetuate this view, but it also conceives agreement as the sole and ultimate goal of deliberation (Palomo Hernández 2025).

In fact, the technosolutionist trend within deliberative circles, including the HM, is not coherent with the very notion of the public sphere, meant to connect the different spaces of a polity through deliberation rather than condensing them. Instead, the combination of consensus-oriented minipublics and technological ‘fixes’ reinforces a minimalistic conception of democracy and deliberation. If technology is supposed to be part of the ‘solution’, what are the ‘problems’? And ‘who’ is meant to take the lead in tackling those problems via technology? This is even more problematic in the context of public spheres that are largely shaped by the ‘algorithmic control of communication flows that is feeding the concentration of market power of the large internet corporations’ (Habermas 2022: 167). Considering that the HM has been developed by Google’s DeepMind, it is problematic to situate Big Tech corporations as both part of the problem and the solution. Quoting Black feminist Audre Lorde, ‘the master’s tools will never dismantle the master’s house’. However, some scholars and practitioners are increasingly building close ties to Big Tech as a way to close the gap between minipublics and mass publics.

For instance, the Stanford Deliberative Democracy Lab (2025) collaborated closely with Big Tech companies to convene a forum using their AI-assisted Stanford Online Deliberation Platform to gather citizens’ views on AI agents. Coherently, the Stanford Deliberative Democracy Lab director, James Fishkin (who owns the registered trademark of ‘deliberative polling’), co-wrote a paper arguing that technology—including AI-assisted moderation—contributed to scaling both the quantity and quality of deliberation (Fishkin et al. 2025). Similarly, Hélène Landemore served in 2023 as an advisor to the Democratic Inputs to AI program at OpenAI, while reproducing an overly optimistic view of the potential uses of AI for deliberation in her recent publications (Landemore 2023, 2024). Interestingly, Landemore herself has acknowledged the limits of private companies investing in democratic innovations, given that “the reality is that these companies are a huge part of the problem to begin with. Their incentives are completely misaligned with the common good” (Giesen 2024). This illustrates how corporate and Global North actors are at the driving seat of technology and participatory innovations, and are likely to continue reproducing our society’s systemic political and economic hierarchies.

Our article suggests that the artificial sense of ‘equality’ present in deliberative minipublics is only worsened through experiments such as the HM, which reinforce the disentangling of democracy from mass politics and undermine the ability to exert collective power through mass membership organisations and the public sphere. Furthermore, it places blind faith on Big Tech to be part of the solution. The alternative to technosolutionism is to orient deliberative democracy towards fostering an agonistic public sphere guided by political contestation rather than a micro ‘representative sample’ or a machine-led pseudo deliberation without citizens. This is not to say that technology is necessarily problematic for democracy. There are interesting recent examples of technology being mobilised in a democratic direction, such as the articulation of worker-owned technologies to strengthen the political power of vulnerable communities (Grohmann 2023). This emphasises that technology can contribute to reimagining and strengthening democracy. But in order to do that, democracy needs to be oriented towards mass politics and collective actors rather than reproducing the narrow conception of deliberation upon which minipublics are built.

Competing Interests

The authors have no competing interests to declare.