Alireza SafarpourArtificial intelligence is no longer only a technical issue, but also a democratic issue, as it is increasingly becoming a powerful force in shaping political systems and democratic governance (Walter, 2024). This article builds on the recently published research by Safarpour (2026) in the ELTE Law Journal, which systematically reviews one decade of scholarship on the relationship between artificial intelligence and democracy. Historically, every major technological shift, from the printing press to the internet, has left its mark on democracy by reshaping how information circulates, how citizens engage, and how governments are held accountable. Therefore, the question is not simply whether AI is good or bad. The real question is how AI is used, who controls it, and whether democratic values remain at the center of this new technological era (Cupać et al., 2024).
Democracy is fundamentally built on the principles of transparency, accountability, citizen participation and fairness. Transparency means that government decisions are communicated clearly and are accessible to the public, which is essential for maintaining trust and legitimacy (Nielsen et al., 2021). Accountability ensures that institutions are answerable for their actions, helping them remain responsive to the needs and concerns of society (Helbing et al., 2023). Genuine opportunities for citizen participation allow individuals to engage in shaping policies, which strengthens the resilience and responsiveness of democratic systems. When AI enters democratic life, it must therefore be judged against these basic democratic principles.
As artificial intelligence and other digital technologies become more embedded in public governance, it is crucial that their integration supports, rather than undermines, democratic foundations. AI refers broadly to technologies that can perform tasks traditionally requiring human intelligence, including machine learning, automation, and data-driven decision-making (Johnson et al., 2022). Today, the concern is that AI is no longer just a tool for e-government but may gradually replace the human role in state decision-making and the exercise of sovereignty (Sætra, 2020; Beckman & Rosenberg, 2022). This makes AI not only a matter of administrative reform, but also a matter of democratic power.
The promise and risk of AI in democratic governance
The positive side of AI is clear. Governments are turning to AI tools to analyse large volumes of data, improve predictions, and streamline administrative processes, all with the aim of making governance more efficient and enabling more informed citizen participation (Rashid & Kausik, 2024). AI-powered platforms can support participatory governance by collecting citizen input and analysing public sentiment, leading to more informed and responsive policymaking (Nechesov & Ruponen, 2024). In this sense, AI may help governments become more capable, faster, and more responsive, but only if human values remain part of the system.
But the negative side is also serious. While artificial intelligence can serve as a powerful tool for enhancing the efficiency of democratic systems, it also raises serious concerns. Among the most pressing issues are the spread of misinformation, algorithmic biases in decision-making, and the growing reach of surveillance technologies (Sun et al., 2024). In other words, the same smart government that can protect rights may also use AI to limit privacy and individual freedoms. These developments show the dual nature of AI: it can improve governance, but it can also weaken democratic control if it is not regulated carefully (Cupać et al., 2024).
One of the most important lessons from the systematic literature review is that AI affects democracy in several connected ways, not only through one single channel. After carefully reviewing and analysing the selected articles, it became clear that artificial intelligence plays a multifaceted role in democracy, shaping it in five key areas: public participation, technocratic decision-making, transparency and accountability, media and political communication, and surveillance and privacy. Therefore, AI should be understood as a force that changes both the tools of democracy and the structure of power itself.
Five areas where AI reshapes democracy
First, AI’s role in public participation presents a paradox: it has the power to enhance democratic engagement while also posing risks to inclusion and fairness. On one hand, AI-driven tools such as chatbots, digital twins, and automated discussion facilitators are expanding access to public debates, improving the quality of deliberation, and making online discussions more structured and inclusive (Hughes, 2017). Democratic oversight is therefore needed to ensure that AI serves the public interest and remains open to citizen participation (Kreps & Jakesch, 2023; Ovadya, 2023). This is important because democracy is not only about elections; it is also about whether citizens can meaningfully participate in shaping the systems that govern them.
Second, AI raises the problem of technocratic decision-making. As reliance on artificial intelligence in governance grows, there is an ever-present concern that the pursuit of algorithmic efficiency might come at the expense of democratic debate and ethical oversight (Sætra, 2020). AI’s impressive computational abilities allow for data-driven policymaking that significantly boosts efficiency; however, this same strength risks reducing politics to technical calculation. This changes the old assumption that governance is always exercised by humans, because part of this authority is now being delegated to AI systems (Beckman & Rosenberg, 2022). This is why democratic governance cannot be reduced to algorithmic efficiency alone; it also needs public reasoning, accountability, and human judgment.
Third, transparency and accountability are central. The integration of advanced AI systems into governance structures has sparked extensive debate about the need for robust transparency and accountability measures that uphold democratic principles (Citino, 2025). One of the central challenges is the so-called “black box” problem, where the opacity of many AI systems makes it difficult to trace decision-making processes. Explainable AI is therefore important because it can make complex systems more understandable to policymakers, citizens, and public institutions (Buhmann & Fieseler, 2023). In easy words, if a public authority uses AI to make or support a decision, citizens should be able to understand why that decision was made, how it was made, and how it can be challenged.
Fourth, AI is changing media and political communication. Artificial intelligence has progressively reshaped the media landscape and transformed political communication in ways that extend far beyond mere technical innovation. Its influence can be seen in how news is produced, curated, and ultimately received by the public, with automated systems sometimes generating “filter bubbles” that restrict access to a variety of viewpoints (Molitorisz, 2024). But these changes also increase the risk of AI-driven misinformation and manipulation. Intelligent personal assistants, search engines, and automated communication tools may influence political perceptions in powerful and sometimes hidden ways (Epstein et al., 2022). This is dangerous because democracy needs a shared public space, reliable information, diverse viewpoints, and political communication that is not secretly manipulated.
Fifth, AI creates new concerns about surveillance and privacy. Recent breakthroughs in artificial intelligence have been quietly transforming how surveillance is conducted, bringing forth a host of challenges that touch every facet of personal and societal life. AI-powered tools, such as facial recognition, predictive analytics, and behavioural profiling, are now employed by both state and corporate entities to monitor public spaces and individual behaviours (Nagy, 2023). In such a context, the state can rely on citizens’ data to organise many public functions, from taxation to insurance and public services. If citizens feel that they are constantly watched, they may become less willing to speak, protest, criticise, or participate.
What democratic societies should do
So, what should we do? Based on the systematic review, four regulatory and ethical principles are especially important: legal accountability and enforcement; transparency, fairness, and ethical AI by design; data governance, privacy, and decentralisation; and democratic oversight, participation, and resilience. These principles show that AI governance should not be only technical; it should also be legal, ethical, and democratic. In simple terms, we need clear rules that prevent discrimination, protect individual freedoms, support responsible innovation, and define responsibility when AI systems cause harm.
Legal accountability and enforcement mean that AI technologies should be governed by legally binding regulations, especially when they affect democratic processes, elections, rights, and public services. Similar to sectors such as pharmaceuticals and environmental protection, AI systems that affect the public interest should be subject to binding rules, testing, and oversight (Nemitz, 2018). High-risk AI applications should be subject to stricter obligations, including impact assessments, independent oversight, and sanctions for misuse. Clear responsibility is also essential, because citizens should know who is accountable when an AI system makes a harmful or unfair decision. In this way, legal accountability can help prevent AI systems from becoming invisible forms of power.
Transparency, fairness, and ethical AI by design mean that AI systems must be developed with inclusivity and user-centered design in mind. They must serve society equitably rather than becoming tools for authoritarian control. From the outset, they must uphold privacy, freedom of expression, the rule of law, fairness, accountability, and non-discrimination (Kuziemski & Misuraca, 2020). This is why regular bias audits, explainability, and human oversight should become basic democratic requirements, not optional technical choices.
Data governance, privacy, and decentralisation mean that AI governance must prevent excessive power concentration and protect individual rights. When control over AI technologies is concentrated in the hands of a few dominant players, whether public or private, it threatens political pluralism and open discourse (Cupać et al., 2024; Nemitz, 2018). Strong data protection laws, competition rules, meaningful consent, and the rule of law are essential parts of democratic AI governance.
Democratic oversight, participation, and resilience mean that citizens, civil society, regulators, researchers, policymakers, academia, industry, and public institutions must all have a role in shaping AI governance. Ensuring democratic oversight is essential to maintaining public trust, protecting electoral integrity, and upholding transparency, inclusivity, and accountability (Duerinckx et al., 2024). Citizen assemblies, participatory AI design, public consultation, and accessible complaint mechanisms can help make AI governance more democratic and less technocratic.
Responsibilities of policymakers, companies, citizens, and researchers
For policymakers, the message is clear: do not wait until AI harms democracy and then react. Regulation should be early, clear, enforceable, and democratic. Policymakers should consider robust AI ethics guidelines, mandatory transparency measures for AI-driven political advertising, and stronger oversight mechanisms to prevent AI-enabled electoral manipulation (Cupać et al., 2024). Democratic societies are beginning to respond, but implementation and enforcement remain crucial.
For technology companies, the message is also clear: democratic societies cannot accept AI systems that are powerful but unaccountable. Companies should explain how their systems work, test them for bias, protect users’ data, and avoid business models that profit from manipulation, polarisation, or surveillance (Kuziemski & Misuraca, 2020). Innovation should continue, but it should not be separated from responsibility, rights, and public trust.
For citizens, the solution is not to reject AI completely, but to become more aware of how it shapes information, political communication, and public decision-making. Citizens need AI literacy, media literacy, and democratic literacy. Democracy in the AI age therefore needs informed citizens, not only strong laws.
For universities and researchers, more empirical research is urgently needed. The review shows that the study of artificial intelligence's impact on democracy is still in its early, exploratory stages. Conceptual articles make up most of the research in this field, with 66 out of 97 studies falling into this category. Research should therefore move from general warnings to concrete evidence about how AI affects participation, equality, rights, and public trust in practice.
Conclusion: keeping democracy at the center of AI governance
In conclusion, AI can help democracy, but it can also harm it. It can support participation, improve public services, and make governance more informed. But it can also spread disinformation, deepen surveillance, strengthen inequality, and move decision-making away from citizens and democratic institutions. In this sense, while AI may support democracy and the rule of law, it can also create new risks for human agency, individual rights, and public oversight. For this reason, the future of AI and democracy should not be left only to engineers, companies, or governments. It must be shaped through democratic debate, legal accountability, public participation, and strong protection of fundamental rights.
_________________________________________________________
References:
Beckman, L., & Rosenberg, J. H. (2022). The democratic inclusion of artificial intelligence? Exploring the patiency, agency and relational conditions for demos membership. Philosophy & Technology, 35(24), 1. https://doi.org/10.1007/s13347-022-00525-3
Buhmann, A., & Fieseler, C. (2023). Deep learning meets deep democracy: Deliberative governance and responsible innovation in artificial intelligence. Business Ethics Quarterly, 33(1), 146–179. https://doi.org/10.1017/beq.2021.42
Citino, Y. M. (2025). Leveraging automated technologies for law-making in Italy: Generative AI and constitutional challenges. Parliamentary Affairs, 78(3), 625–647. https://doi.org/10.1093/pa/gsae040
Cupać, J., Schopmans, H., & Tuncer-Ebetürk, İ. (2024). Democratization in the age of artificial intelligence: Introduction to the special issue. Democratization, 31(5), 899–905. https://doi.org/10.1080/13510347.2024.2338852
Duerinckx, A., Veeckman, C., Verstraelen, K., Singh, N., Van Laer, J., Vaes, M., Vandooren, C., & Duysburgh, P. (2024). Co-creating artificial intelligence: Designing and enhancing democratic AI solutions through citizen science. Citizen Science: Theory and Practice, 9(1), 43, 1–15. https://doi.org/10.5334/cstp.732
Epstein, R., Lee, V., Mohr, R., & Zankich, V. R. (2022). The answer bot effect (ABE): A powerful new form of influence made possible by intelligent personal assistants and search engines. PLOS ONE, 17(6), e0268081. https://doi.org/10.1371/journal.pone.0268081
Helbing, D., et al. (2023). Democracy by design: Perspectives for digitally assisted, participatory upgrades of society. Journal of Computational Science, 71, 102061.
Hughes, J. (2017). Algorithms and posthuman governance. Journal of Posthuman Studies, 1(2), 166–184. https://doi.org/10.5325/jpoststud.1.2.0166
Johnson, P. C., et al. (2022). Digital innovation and the effects of artificial intelligence on firms’ research and development: Automation or augmentation, exploration or exploitation? Technological Forecasting and Social Change, 179, 121636.
Kreps, S., & Jakesch, M. (2023). Can AI communication tools increase legislative responsiveness and trust in democratic institutions? Government Information Quarterly, 40(3), 101829. https://doi.org/10.1016/j.giq.2023.101829
Kuziemski, M., & Misuraca, G. (2020). AI governance in the public sector: Three tales from the frontiers of automated decision-making in democratic settings. Telecommunications Policy, 44(6), 101976. https://doi.org/10.1016/j.telpol.2020.101976
Molitorisz, S. (2024). A legal cure for news choice overload: Regulating algorithms and AI with “light patterns” to foster autonomy and democracy. Policy & Internet, 16(3), 643–660. https://doi.org/10.1002/poi3.412
Nagy, N. (2023). “Humanity’s new frontier”: Human rights implications of artificial intelligence and new technologies. Hungarian Journal of Legal Studies, 64(2), 236–267. https://doi.org/10.1556/2052.2023.00481
Nechesov, A., & Ruponen, J. (2024). Empowering government efficiency through civic intelligence: Merging artificial intelligence and blockchain for smart citizen proposals. Technologies, 12, 271.
Nemitz, P. (2018). Constitutional democracy and technology in the age of artificial intelligence. Philosophical Transactions of the Royal Society A, 376(2133), 20180089. https://doi.org/10.1098/rsta.2018.0089
Nielsen, J., Eckstein, L., Nicol, D., & Stewart, C. (2021). Integrating public participation, transparency and accountability into governance of marketing authorisation for genome editing products. Frontiers in Political Science, 3, 747838.
Ovadya, A. (2023). Reimagining democracy for AI. Journal of Democracy, 34(4), 162–170. https://doi.org/10.1353/jod.2023.a907697
Rashid, A. B., & Kausik, M. A. K. (2024). AI revolutionizing industries worldwide: A comprehensive overview of its diverse applications. Hybrid Advances, 7, 100277.
Safarpour, A. (2026). Artificial intelligence and democracy: A systematic literature review covering one decade. ELTE Law Journal, 2026(1), 85–118. https://doi.org/10.54148/ELTELJ.2026.1.85
Sætra, H. S. (2020). A shallow defence of a technocracy of artificial intelligence: Examining the political harms of algorithmic governance in the domain of government. Technology in Society, 62, 101283. https://doi.org/10.1016/j.techsoc.2020.101283
Sun, W., McFaddin, S., Tran, L. H., et al. (2024). PresAIse, a prescriptive AI solution for enterprise. INFOR: Information Systems and Operational Research, 62(4), 629–632.
Walter, Y. (2024). Managing the race to the moon: Global policy and governance in artificial intelligence regulation—A contemporary overview and an analysis of socioeconomic consequences. Discover Artificial Intelligence, 4, 14, 1–25. https://doi.org/10.1007/s44163-024-00109-4
_________________________________________________________
The views expressed above belong to the speakers and do not necessarily represent the views of the Centre for Social Sciences.