One of the most common empty phrases about artificial intelligence (AI) is that it offers opportunities and challenges. Yet, AI can improve citizen participation when AI is not corrupted by large corporations using AI for profit. Interestingly, in May 2026, the New Yorker asked, “The A.I. Industry Is Booming. When Will It Actually Make Money?”

To steer the commercially — and seemingly not yet profitable — defined so-called “landscape of AI” that might be used for participation, the democratic essentials need to be considered away from so-called “emerging trends” of AI. As if such “emerging trends” fall out of the sky. What also does not fall out of the sky is deliberative democracy (Habermas and Rawls).

Deliberative democracy centers on reasoned public discussion as the core of legitimate political decision-making. In standard democracy, legitimacy comes from simply counting votes. Deliberative democracy argues this isn’t enough. It holds that political decisions are legitimate only if they emerge from a process of open, reasoned discussion among free and equal citizens.

Given these precursors, AI might still offer a wide range of opportunities in supporting democratic participation. Democratic AI tools might even assist democratic governments and locals when addressing some of the challenges they face in conducting deliberative participation. Of course, in such scenarios, corporations are never mentioned, as they are hierarchical, semi-dictatorial, and authoritarian institutions.

This frames the AI example away from corporations — democracy should not bother places where many people spend eight hours per day! Having saved corporations from the evils of democracy, AI support for participation can be located (or better, asphyxiated) in — at least potentially — seven ways:

1. By improving the design and accessibility of citizen participation.
2. By enabling the implementation of democratic participation.
3. By interpreting the goals and outcomes of democratic participation.
4. By communicating the outcomes of democratic participation.
5. By connecting people to the broader public.
6. By bypassing corporate media and their power to manipulate agendas.
7. By reducing the costs of citizen participation processes.

A recent OECD report on AI proposes a typology — away from corporations — to “navigate” (better: channel) these opportunities to the public domain. The OECD‘s typology identifies nine types of applications of AI tools to support and improve democratic participation:

1. Information development
2. Sense-making
3. Translation
4. Transcription
5. Virtual assistance
6. Moderation
7. Facilitation
8. Simulation
9. Architecture

These AI functions assist government officials, practitioners, and citizens in the various tasks and activities required to implement and take part in democratic participatory and deliberative processes.

In some cases, these AI tools might even enable new forms of democratic participation altogether, or at scales not feasible without the use of AI. For each application, the OECD‘s typology explores the relevant uses of AI tools in, for example:

• the front office — the citizen-government interface; and
• the back office of government — governments’ internal operations.

Finally, each application is analyzed against its relevance in addressing the existing gaps and challenges to achieving meaningful and effective participation. For example, AI-powered sense-making, virtual assistance (e.g., personalized support to participants through conversational interfaces), and translation tools stand out as the most widespread.

Among the different types of citizen participation processes, deliberative processes constitute a lively field of experimentation with AI tools, in particular to improve live facilitation activities and to enable deliberation at scale.

Beyond citizen participation processes in democracy, AI presents relevant opportunities to support the implementation of broader open government policies and practices, including access to information, government communication, and the protection of civic space.

Future developments in AI might encourage a reflection on more cutting-edge and emerging practices at the “intersection” (now an overused buzzword) of AI and participation, such as possible uses of AI agents in citizen participation processes and human-AI policymaking.

Many governments around the world are already using AI tools to design, manipulate, and implement citizen participation processes taking place at all levels of governance — local, state, and federal, for example.

However, like other types of AI efforts, the adoption of AI mostly occurs on an ad hoc basis or through pilot initiatives — regularly from above, not from below. While some local governments drive this trend, a growing number of national governments are developing coordinated strategies or tools to integrate AI into their participatory processes, knowing that such uses of AI in citizen participation entail risks and challenges, such as the risk that real democracy might take place and that this might challenge established forms of domination.

Building on the analysis carried out in the OECD’s report, this piece distinguishes between risks and challenges and focuses on their respective implications in the use of AI applied to citizen participation. With respect to risks, these can be found in:

1. Morality: ethical risks, such as inadequate, manipulated, and skewed data in AI systems. This leads to “inaccurate” outcomes that stabilize existing forms of domination — politicians over people.
2. Transparency: the lack of transparency and explainability of some AI tools can further be used to manipulate the entire process.
3. Democratic Values: this allows the users of AI to present AI as value-neutral while enhancing their agenda, or even an anti-democratic agenda. Beyond that, AI can simply be used for “agenda setting” purposes, thereby bypassing people and democratic processes.
4. Public Domain: worse, the misuses of AI can undermine civic space or manipulate it for unsavory purposes.
5. Operational Risks: these can come from an overreliance on AI systems, making up key aspects of governing and presenting them as facts.
6. Hallucination: beyond that are the delusionsof AI — when AI has gone astray and makes up things.
7. The Private Sphere: Next, there are privacy and data concerns— as always with an intrusive system like AI.
8. Threads: There are also cyber threatsthat can increase when AI is used and when AI attacks other systems.
9. Trust: All of these result in the manipulation of participation and further erode public trust in democracy: as anarchists like to say, if democracy would change things, it would have been made illegal!
10. Exclusions: next, there are segregation risks. This includes the deliberate appropriation or fabrication of a digital language that is societally divisive.
11. Domination: for established forms of domination, even when framed as democracy, there is always the risk of public resistance against the domination of a particular group, corporations, or even against capitalism.
12. Paralyzing Democracy: The application of AI in democracy might also lead to indecision — under the illusion that AI is doing it! This can later be framed as “missed opportunities.”

Beyond these twelve risks for people and democracy, governments also face organizational and institutional challenges. These are related to the integration of AI into government operations, which have to be done in such a way that people get the feeling that AI enhances democracy. In reality, it can manipulate the entire process toward the stabilization of systems of domination.

One challenge for people is that AI can exploit existing gaps in skills and literacy, which, in turn, further stabilizes domination. It can create a much-encouraged and convenient overreliance on AI systems that further enhances the power of a dominating group — for example, C. Wright Mills’ Power Elite.

The introduction of AI might lead to inadequate investment in other areas, like schools and hospitals, for example. Under the cover of using AI to enhance democracy, it might lead to a scaling down of governmental practices in other areas. On Lenin‘s famous question of “what is to be done?” at least twelve recommendations can be extrapolated:

1. Local participants must ensure that the actions taken by local governments create and enhance democracy and conditions that benefit people.
2. The opportunities offered by participatory AI tools need to improve citizen participation.
3. AI processes need to mitigate the risks of enhancing and further cementing asymmetrical power structures.
4. People need to establish “guardrails” to anticipate and deal with the risks associated with AI adoption in participation.
5. AI must be adopted fairly, morally, and transparently. AI systems in governance cannot be left without supervision and control by a democratic citizenry.
6. Such a democratic community needs to ensure that even supposedly democratic AI systems are compliant with the democratic values that define a modern society.
7. The protection of civic space needs to be enhanced, not damaged, by AI.
8. With the right supervision, AI might assist in closing the data divide — not boosting it further.
9. Democratic citizens should seriously consider adopting open-source AI. Beyond that, locals who introduce AI should ensure they retain the ability to switch it off.
10. People should act to prevent the exclusiveness of AI by, for example, offering low-tech alternatives for citizens.
11. Democratic AI would also mean improving AI literacy for all.
12. Locals must ensure that governments are transparent in the implementation and use of AI.

With democratic oversight, AI can foster engagement with citizens while shaping and governing AI systems for democratic participation. AI can also work toward more meaningful and impactful citizen participation by addressing asymmetrical institutional relationships of power.

Beyond all this, locals should realize that AI can never solve issues of participation. This requires democracy — not AI. Meanwhile, participants in democratic processes should continue to work toward improving the quality of democracy and the accessibility of participation processes to citizens. They need to establish democratic mechanisms to ensure that citizens’ efforts – whether with or without AI – have a clear impact on government policy and decision-making.

In all this, AI is no more than a supporting tool to further improve and enhance meaningful, impactful, and democratic citizen participation. Finally, citizens participating in AI processes need to weigh the potential benefits of AI against the potential dangers of AI — dangers that lie in further cementing structures of domination rather than assisting people in the process of democratizing these structures and institutions.

On the other hand, people can leverage deliberative-democratic participation in the design, introduction, deployment, monitoring, overseeing, and democratic control of AI systems. Looking ahead, any democratic community needs to consider how democratic institutions and deliberative participation can thrive once AI has become part of democratic institutions.

---