Corruption is viewed as one of the major concerns of modern-day democracies – a chronic lack of transparency. Accountability is distributed across institutional arrangements, processes, and networks, making it hard to identify who bears responsibility for what. The quick pace of media production makes scandals emerge and subside without any deeper knowledge on the part of the audience. This leads to the situation where people find themselves increasingly incapable of comprehending the ways in which political power is organized and practiced.
The emergence of artificial intelligence into this scenario is a consequence of the self-consciousness of democracy itself regarding its unreadability rather than an external factor influencing its development. Often seen as either a source of authoritarianism or a method to restore democratic governance, AI has neither inherently liberating nor dominating nature. As such, it cannot serve as a political actor independently but can only help reveal pre-existing power structures. Hence, its importance should be sought in making political processes transparent rather than in overthrowing any particular system of government.
Furthermore, AI may alter political trust beyond merely enhancing its efficiency. In the contemporary context, distrust stems from an inter-generational divide where older people view politics as unstable and overly fluid, while the younger generation sees it as institutionally static. This double-sided nature makes politics appear either unpredictable or stuck in its structures, both of which weaken political trust. The proper use of AI, instead of making politics simpler, can play a crucial role in ensuring greater transparency regarding its structural aspects.
The idea here may consist of conceiving AI as a public political database where all data related to governance are organized and represented in a structured form. This approach will be based on John Rawls’ “veil of ignorance” in which politics are presented in a non-institutionalized and non-stakeholdered form. The proposed political database will include maps of connections between political stakeholders, business companies, and governmental institutions as well as records of decision-making within political institutions.
More importantly, such an approach would have to consider indirect methods of power as well. As Bachrach and Baratz state, apart from the influence of decisions made, power is manifested via the process of decision-making by choosing which questions may be
put on the agenda. Thus, an AI system for transparency would have the task of detecting any patterns of exclusion, delays, or procedural obstruction of political decisions.
In contrast to Lukes’ theory of power which focuses on the manipulation of beliefs, this concept would exclude any discussion of ideology and beliefs per se. Instead, it would try to use purely behavioral data and interaction between political institutions and procedures to understand how politics works, rather than what it means.
What lies at the heart of the proposed system is the problem of diminishing political accountability. In more established democratic regimes, this results in diffusion of responsibility due to long chains of institutional interaction. By comparison, in more underdeveloped democracy, it is due to corruption, fragility of institutions, and the fragmentation of political memory. Nevertheless, transparency remains insufficient due to lack of coherence, rather than absence of information.
A map-driven approach to transparency powered by AI technologies can be regarded as a constantly updated network of political connections, reflecting decision-making processes at different moments in time. It does not imply viewing politics as sporadic events but instead tries to restore the evolving web of relationships that make up the political world. Such a model bears striking similarity to strategic cartography of governance.
Access will also play an essential role here. Even though the system itself will be located online and be constantly updated, it may have other versions, such as radio broadcasts or printed briefs, which would allow for accessibility among those who are less actively
involved in Internet-based social activities. The ultimate aim is full democratic transparency.
What is especially valuable about such an approach is the fact that it should not function as any sort of interpretation of political phenomena. Instead, it will serve as a transparent index, very much like some governance indices or media freedom indices.
The importance of such a model can be seen clearly upon looking at some real-life applications of AI in politics. In the US, there are examples of AI-powered content creation playing roles in some recent political races. For instance, Donald Trump made use of several images created using artificial intelligence during the 2024 political cycle. This involved the use of artificial imagery showing support for Trump from various personalities and flattering images of him, alongside other images which attempted to discredit political opponents like Kamala Harris.
In Hungary, AI is more focused on creating an emotional frame in the politics of the country. In this case, AI plays a more instrumental role in affecting the emotions of
citizens rather than building individualized brands for politicians. This is because AI is mainly employed in stirring feelings based on issues of nationalism and patriotism.
Taiwan offers another form of a hybrid where disinformation based on AI is witnessed while at the same time creating civic technology such as vTaiwan for dialogue. Taiwan passed the AI Basic Act in 2026, indicating the adoption of AI.
In all of these examples, a shared thread becomes evident: AI is neither a neutral nor an autonomous driver of events; rather, it is a versatile tool that is designed by institutions’ purposes. It is utilized by political players to persuade citizens; governments harness it for narrative-building purposes; and deliberation processes employ AI to encourage participation. And so, the crux of the matter lies in how we will govern its visibility and legibility.
The critical aspect that determines AI’s influence on democracy is the specific institutional contexts where it is introduced. AI per se does not rework democratic processes, but amplifies their current state, making them more understandable or incomprehensible depending on circumstances. The question of importance here is how democracies will design their accessibilities and legibilities.