Artificial Intelligence and Digital Political Communication: A Framework for Understanding Democratic Resilience and Vulnerability ()
1. Introduction
Every phase of technological innovation has inevitably transformed the ways in which individuals communicate, build relationships, and interpret the reality around them. From the spread of the printing press to the advent of the internet, communication tools have always represented forces capable of reshaping the social and cultural dynamics of communities (McLuhan, 1994). The digital revolution has further accelerated this process, introducing new spaces of interaction in which the production and circulation of information occur with unprecedented speed and reach. Within this environment, characterized by the growing centrality of data and algorithms, the ways in which public narratives are constructed have also undergone profound changes (Van Dijck et al., 2018). Communication is no longer limited to the simple transmission of content but has become a dynamic process in which personalization, automation, and analytical capabilities play an increasingly significant role. Since the emergence of social media, new forms of political participation and unprecedented spaces for the production, circulation, and social influence within the digital public sphere have developed. The progressive rise of online platforms has reshaped the communicative practices of political actors, facilitating a transition from predominantly one-way communication models, based on institutional websites and blogs, toward more immediate, participatory, and personalized forms of interaction through social environments (Chadwick, 2017). Within this context of continuous technological transformation, the communication strategies adopted during electoral campaigns have also undergone significant evolution, adapting to new digital infrastructures and changing patterns of political information consumption. The introduction of artificial intelligence (AI) into communication processes has further accelerated these transformations, generating social and political phenomena unprecedented in terms of both scale and speed of diffusion. These technologies have produced significant effects on the dynamics of contemporary society, particularly affecting the mechanisms through which informational content is produced, distributed, and interpreted (Papathanassopoulos & Giannouli, 2025). Indeed, AI has altered the ways in which contemporary events are represented, filtered, and perceived within the public sphere, introducing new levels of automation in the management of communication flows. At the same time, the digital revolution has profoundly transformed forms of social interaction, enabling individuals from different cultural and linguistic backgrounds to communicate through advanced technological tools, such as chatbots, virtual assistants, and conversational systems based on machine learning. These instruments have expanded opportunities for access and participation, while also introducing new challenges concerning algorithmic transparency, personal data management, and the quality of information available online (Zuboff, 2019). Within the field of political communication, the integration of artificial intelligence has substantially transformed the ways in which electoral messages are designed, distributed, and received, particularly affecting strategies of political campaigning and voter mobilization (Strömbäck & Kiousis, 2014). During electoral campaigns, intelligent systems are increasingly employed to optimize internal decision-making processes, analyze large volumes of data, and improve the ability of political actors to reach specific segments of the population (Kugler, 2018). Through techniques such as micro-targeting, user profiling, and automated experimentation with communicative content, campaigns can be directed toward specific categories of voters, who are exposed to personalized messages through digital platforms (Battista, 2026a). The use of these tools enables political actors to refine their persuasive strategies by improving their ability to identify the needs, interests, and orientations of the electorate. From this perspective, AI represents a strategic resource capable of making political communication more adaptive and data-driven, encouraging innovative forms of civic engagement and democratic participation (Battista, 2025). However, these opportunities coexist with significant concerns related to the improper use of algorithmic technologies and the possibility that they may be employed to distort information processes and influence voters’ behaviour (Bender, 2022). The democratic implications of artificial intelligence become particularly relevant when considering the ability of these systems to produce manipulative content, simulate identities, and undermine the distinction between authentic and artificial communication. Algorithms can be employed to disseminate misleading information, impersonate public figures, or conceal the origin of content directed toward voters (Kreps & Kriner, 2023). Among the most debated practices are the production of realistic deepfakes, the automated generation of political messages, and the creation of apparently spontaneous campaigns that are coordinated through artificial systems (Battista, 2024). These phenomena become even more concerning when considering the rapid improvement of generative technologies in their capacity for imitation and persuasion, increasing the risk of altering public perceptions of reality and weakening trust in democratic institutions (Krishnarajan, 2023). Considering these transformations, this article aims to examine how artificial intelligence is reshaping contemporary political narratives, focusing on the consequences arising from the automation of electoral communication, the digital substitution of candidates’ identities, and the manipulation of public discourse. The analysis is developed through a series of paradigmatic cases concerning the use of AI in recent electoral campaigns, with the objective of highlighting both the opportunities offered by these technologies and the risks associated with their integration into democratic systems.
2. Changes in Digital Political Communication
In recent years, the media system has undergone a profound structural transformation, characterized by the progressive decline of the role of traditional media and the emergence of digital platforms as the primary environments for the production, distribution, and consumption of political information (Kleinnijenhuis et al., 2019). Prior to the widespread adoption of online communication strategies, electoral campaigns relied predominantly on traditional mass media, such as television and radio, through a largely vertical and generalized communication model, in which the same message was directed indiscriminately to different audiences. With the rise of social media, this paradigm has gradually been replaced by a model characterized by greater interactivity, participation, and personalization. Unlike traditional media, social platforms enable users to assume an active role in the processes of producing, circulating, and interpreting informational content. The digital citizen is no longer merely a passive recipient of political messages but becomes an integral part of online communities in which individuals contribute to the construction and dissemination of public narratives (Bode, 2016). Moreover, within these digital environments, the constraints imposed by geographical distance are progressively reduced, allowing individuals to share opinions, participate in public debates, and engage in political mobilization in real time on a global scale (Johann, 2022). Within this new communication ecosystem, political communication has undergone a radical transformation, altering the ways in which candidates, political parties, and voters interact with one another. Digital platforms have introduced more immediate and direct forms of interaction, reduced the role of traditional intermediaries and fostered a more personalized relationship between political representatives and citizens (Kovarik, 2025). Contemporary electoral campaigns increasingly rely on social media for the dissemination of political messages, the mobilization of supporters, the organization of events, and the construction of electoral consensus. From this perspective, new media should be understood not only as tools for political information but also as spaces in which collective identities, social movements, and new forms of democratic participation are formed (Bennett & Segerberg, 2023). Users actively engage in public debate through digital communities characterized by the sharing of common values, interests, and political orientations, transforming social media into environments capable of both reflecting and influencing the transformations of contemporary political participation. At the same time, electoral campaigns have progressively assumed a dimension increasingly oriented toward data collection and analysis, giving rise to what several scholars have defined as the “fourth era of political campaigns”. During this phase, the identification of potential voters takes place through sophisticated systems of behavioral analysis, profiling tools, and personalized communication strategies aimed at disseminating messages constructed according to users’ individual characteristics. This evolution has resulted in a shift from campaigns targeting broad segments of the electorate to campaigns characterized by a high degree of communicative granularity. Within this process, social platforms play a central role in the selection and distribution of the political content to which users are exposed. Through automated algorithmic systems, these digital infrastructures determine which content receives greater visibility, which users can be reached by political advertising campaigns, and what costs advertisers must incur to target specific groups of voters (Crain & Nadler, 2019). This practice, known as political microtargeting, involves the use of personal data and behavioral information to deliver highly personalized political messages based on variables such as age, gender, interests, cultural orientations, and individuals’ online activities (Stark et al., 2020). The primary objective of these strategies is to increase the persuasive effectiveness of political communication by offering content that is more relevant to recipients and by enhancing the likelihood of voter engagement. In this context, every digital interaction from visiting a page to reacting to content or participating in an online discussion can be transformed into an indicator useful for constructing individualized political profiles (Duberry, 2023). However, the use of these techniques raises numerous concerns, as algorithmic systems may produce inaccurate classifications or incorporate biases, generating distorted representations of individuals and their political orientations (Shin, 2024). A further problematic element concerns the functioning of recommendation algorithms employed by digital platforms. To maximize user engagement, these systems often prioritize content consistent with individuals’ previously expressed preferences and interests, contributing to the creation of highly personalized information environments. Since algorithms do not necessarily assess the quality or reliability of the information being disseminated, users may become progressively exposed to homogeneous communicative contexts, in which existing political beliefs are reinforced and opportunities for engagement with alternative perspectives are reduced (Dahlgren, 2021). In this sense, artificial intelligence does not merely perform a function of selecting available content but actively contributes to the construction of the informational context within which individuals develop their perceptions of political reality. This dynamic may foster phenomena such as polarization, misinformation, and fragmentation of the public sphere, negatively affecting the quality of democratic debate (Battista, 2026b).
3. Methodological Note
Although it primarily takes the form of a theoretical-conceptual contribution, this article integrates theoretical reflection through a qualitative comparison of the cases of Indonesia and Taiwan region, selected according to a logic of theoretical case selection, consistently with the methodological approaches adopted in comparative social science research. The choice of these contexts does not derive from criteria of statistical representativeness, but rather from the aim of examining two institutional configurations characterized by different models of artificial intelligence governance applied to political communication and electoral processes. Although both countries have experienced an increasing integration of AI-based tools within electoral campaigns and digital information ecosystems, they display significantly different regulatory frameworks, levels of institutional maturity, and governance strategies. This makes them particularly suitable for identifying the conditions that foster democratic resilience or increase democratic vulnerability. In this contribution, democratic resilience is understood as the capacity of a democratic system to adapt to the transformations introduced by artificial intelligence while preserving the integrity of electoral competition, institutional transparency, and the quality of public deliberation. It does not imply the absence of risks; rather, it refers to the existence of institutional, regulatory, and social mechanisms capable of preventing, absorbing, and mitigating the potentially destabilizing effects arising from the use of AI in political communication. Conversely, democratic vulnerability identifies the degree to which institutions and the public sphere are exposed to the disruptive effects produced by algorithmic manipulation, the automated dissemination of disinformation, the opacity of recommendation systems, and the progressive erosion of citizens’ trust. Throughout the analysis, these concepts are examined through five fundamental dimensions: institutional oversight capacity, platform governance, technological transparency, citizens’ digital literacy, and levels of public trust. The interaction among these elements determines whether artificial intelligence becomes a factor of democratic strengthening or, conversely, contributes to the weakening of democratic quality. The interpretative framework developed in this study assumes that the democratic effects of artificial intelligence are, therefore, the result of the interaction among these closely interconnected dimensions. The first concerns technological capability, namely the level of sophistication of AI systems employed in the production, distribution, and personalization of political content. The second dimension is institutional oversight, understood as the capacity of public institutions to regulate and monitor the use of artificial intelligence through regulatory instruments and accountability mechanisms. This is complemented by platform governance, which encompasses content moderation policies, algorithmic transparency, and practices for the identification of synthetic content. The final two dimensions are represented by civic capacity, referring to citizens’ level of digital literacy and informational resilience, and public trust, understood as the degree of legitimacy attributed to democratic institutions and electoral processes. Democratic resilience emerges when these dimensions operate in a complementary and mutually reinforcing manner, whereas imbalances or deficiencies in one or more areas significantly increase the vulnerability of the democratic system. The analysis is based on a triangulation of sources consisting of peer-reviewed international scientific literature, institutional documentation, reports produced by international organizations, policy papers, governmental documents, and publications developed by research centers specializing in digital governance and artificial intelligence. This plurality of sources enables the reconstruction of the regulatory, technological, and political environments within which AI is implemented in electoral communication, while providing a multidimensional perspective capable of integrating technological, institutional, and socio-political dimensions. More specifically, comparative analysis examines the role played by public institutions in regulating artificial intelligence, the policies adopted by digital platforms regarding content moderation and algorithmic transparency, the strategies developed to counter disinformation and synthetic content, as well as the contribution of civil society and digital literacy initiatives in strengthening citizens’ critical capacities. Furthermore, the analysis focuses on how these dimensions interact in shaping the balance between technological innovation and the protection of democratic principles, with particular attention to the relationship between institutional capacity, platform governance, civic participation, and public trust. From this perspective, the cases of Indonesia and Taiwan region are not considered merely as descriptive examples, but rather as empirical contexts through which the conceptual categories developed within the proposed theoretical framework can be examined and critically assessed. Therefore, the objective of the comparison is not to generate statistical generalizations or establish universally valid causal relationships, but rather to identify recurring configurations, institutional dynamics, and interpretative mechanisms that allow for a deeper understanding of how different governance models influence the capacity of democratic systems to adapt to the transformations introduced by artificial intelligence, either strengthening or undermining the resilience of democratic institutions and the digital public sphere.
4. Challenges and Dilemmas
The exploration of recent electoral processes highlights how artificial intelligence does not act exclusively as a tool for informational distortion or the manipulation of public consensus but also represents a strategic resource through which political actors redefine their methods of communication with the electorate. A particularly significant case is the Indonesian presidential election of February 2024, during which the use of generative AI-based technologies demonstrated how such tools can be employed to construct new forms of political propaganda capable of combining personalization, emotional engagement, and automated content production (Wihardja et al., 2025). The Indonesian context is particularly relevant because it represents one of the largest democracies in the world, characterized by a young population deeply connected to digital environments, in which social platforms have assumed a central role in shaping public opinion. During the 2024 electoral campaign, candidates made extensive use of digital tools to reach primarily younger segments of the electorate, which constitute a decisive component of the voting population. Within this scenario, artificial intelligence was mainly employed to produce synthetic images, personalized audiovisual content, and digital representations of candidates aimed at strengthening their popularity online. The most debated case involved the winning candidate, Prabowo Subianto, whose campaign utilized AI-generated images portraying the candidate through a “younger” and more accessible version of his public persona, characterized by graphic elements inspired by popular culture and the communication styles typical of social media platforms. This communication strategy produced a significant transformation in the construction of the candidate’s political identity, reshaping the traditional image of a former general associated with a more institutional form of politics into a digital representation closer to younger users. Through avatars, artificially generated illustrations, and easily shareable content, the electoral campaign exploited the capabilities of generative AI to create a more emotional and immediate narrative, demonstrating how contemporary political competition no longer develops solely at the level of ideas and policy programs, but also through the ability to control the symbolic representation of candidates within digital spaces (Nugroho, 2026). However, the use of these technologies has raised numerous questions concerning the relationship between political authenticity and the manipulation of public image. Although AI-generated content was not necessarily intended to disseminate false information, it contributed to the creation of an artificial version of the candidate, constructed through aesthetic and communicative elements designed to enhance users’ emotional support. This dynamic highlights a central aspect of the impact of artificial intelligence on contemporary politics: the risk does not concern exclusively the production of fake news, but also the gradual replacement of authentic political communication with synthetic representations optimized by algorithms. A further problematic element that emerged during the Indonesian elections concerns the spread of manipulated content through platforms such as TikTok, Facebook, and X, where artificially generated images and videos reached millions of users within extremely short periods of time. The speed at which such materials circulated made the work of verification systems and independent fact-checking organizations increasingly complex, highlighting one of the main challenges posed by generative AI: its ability to produce plausible content at a speed that exceeds the capacity of democratic institutions to monitor and regulate it (Chesney & Citron, 2019). In particular, the Indonesian case demonstrates how artificial intelligence can contribute to the transformation of electoral propaganda through a progressive aestheticization of politics. The creation of attractive and easily shareable images enabled candidates to reach segments of the electorate that are normally less interested in traditional politics, while simultaneously encouraging a form of communication increasingly based on appearance and emotional identification rather than on the critical discussion of political proposals. This phenomenon is particularly relevant in contexts characterized by high social media penetration, where users’ attention represents a crucial competitive resource. Despite these challenges, the Indonesian case also demonstrates that AI should not be interpreted exclusively as a threat to democratic systems. The transparent and regulated use of these technologies may in fact provide new opportunities to improve political communication, facilitate citizen engagement, and make electoral messages more accessible. The central issue therefore lies in the existence of adequate oversight mechanisms capable of ensuring that technological innovation does not become a mechanism for manipulating public opinion (Sudibyo & Nugroho, 2026). Ultimately, the 2024 Indonesian elections demonstrate how artificial intelligence represents a new frontier of political communication, in which the boundary between innovation and manipulation is becoming increasingly blurred. The Indonesian experience shows that the main democratic risk does not consist solely in the spread of false information, but rather in the possibility that artificial tools may become capable of constructing alternative political realities, in which the perceived image of a candidate prevails over the critical evaluation of their positions and political program.
5. AI as a Tool for Strengthening Democratic Functions in Electoral Campaigns
Evidence from a variety of political contexts suggests that the use of artificial intelligence does not automatically produce negative consequences for the quality of democracy, as its effects vary depending on the political and social environment in which it is deployed (Font, 2020). While in many contexts AI has contributed to phenomena such as polarization and disinformation, the case of Taiwan region demonstrates that, starting from virtually identical technologies, it is possible to strengthen democratic resilience through practices of active citizenship. The Taiwan region STPI case illustrates how generative AI (GenAI) can become a supporting infrastructure for public policymaking. The Retrieval-Augmented Generation (RAG) model makes it possible to connect sources, documents, and policy decisions, providing valuable lessons in governance, institutional capacity, public procurement, and administrative quality. This suggests that the primary challenge lies not in technological innovation itself, but rather in the ways in which societies, institutions, and digital platforms choose to integrate these new tools into communication and governance processes. In a context characterized by the growing disengagement of citizens from politics, AI may represent an important opportunity to renew electoral campaigns and reshape the relationship between political actors and voters (Nemitz, 2018). Taiwan region’s experience represents one of the most significant examples of the use of artificial intelligence in support of democratic resilience, highlighting how the effects of technology largely depend on the ways in which it is governed. In particular, the initiative developed by the Science and Technology Policy Research and Information Center (STPI) demonstrates the potential of Retrieval-Augmented Generation (RAG)-based solutions in improving the reliability of information made available to citizens and in countering the spread of manipulative content during electoral processes. By integrating advanced language models with continuously updated institutional knowledge bases, the system can prioritize verifiable information and reduce the risk of informational hallucinations typically associated with traditional generative models. This approach promotes greater transparency within the information ecosystem, strengthens public trust, and supports more informed political participation. The case therefore demonstrates that the capacity of artificial intelligence to contribute to democratic strengthening does not depend exclusively on the level of technological innovation achieved, but rather on the quality of institutions, the cooperation between the public sector and civil society, and the adoption of effective models of democratic governance. Through chatbots, virtual assistants, and automated communication systems, candidates can disseminate political information more rapidly and directly, making political content more accessible to the public. This strategy appears particularly effective for younger generations, who are increasingly accustomed to receiving information through digital platforms and personalized communication formats. Moreover, these technologies offer the possibility of modernizing electoral communication by making it clearer and more accessible. Through automated translation, text summarization, and targeted content selection, AI can overcome significant information barriers, thereby expanding opportunities for citizens to participate in public debate (Jungherr, 2023). Such tools are especially valuable in societies characterized by linguistic, cultural, or generational diversity, where they can simplify the understanding of political proposals and encourage greater civic engagement (Saud et al., 2023). For these reasons, AI does not merely reshape campaign strategies but can also significantly influence the way citizens interact with democratic institutions. Owing to their ability to process vast amounts of data, political parties can gain a more detailed understanding of voters’ needs, identifying concerns and interests that would be difficult to detect through more traditional forms of political campaigning (Denton Jr. et al., 2023). An additional important contribution of artificial intelligence lies in its potential to counter disinformation. As demonstrated by the case, algorithmically assisted fact-checking systems can substantially reduce the spread of manipulated content by verifying questionable materials within a very short time frame (Baptista et al., 2026). In an environment increasingly threatened by deepfakes, the development of rapid and effective mechanisms for detecting and countering manipulation has become essential. Taiwan region’s experience reinforces the idea that technology, when combined with digital literacy, institutional trust, and cooperation between citizens, can become a crucial instrument for safeguarding democracy. Furthermore, these systems can reduce the virality of fake news, ensuring that misleading content remains on social media platforms for progressively shorter periods (Berrondo-Otermin & Sarasa-Cabezuelo, 2023). Another significant aspect of AI integration into electoral campaigns concerns the evolving relationship between political leaders and voters. Operating within digital platforms characterized by continuous and immediate communication enables candidates to monitor public reactions to specific political issues in real time (Battista & Cervi, 2025). This development transforms traditional forms of political communication into more dynamic processes based on continuous flows of content and feedback. In some cases, such immediacy may foster a stronger perception of closeness between political representatives and citizens, thereby reinforcing citizens’ sense of participation in public life. At the same time, the ability of algorithms to identify topics receiving widespread online attention allows campaigns to adapt their language to current communication needs, making political debate more responsive and interactive. Despite these opportunities, however, several significant challenges remain. When voters receive different political messages tailored to their personal characteristics, ideological preferences, digital behaviors, or emotional vulnerabilities, democratic deliberation risks losing its collective dimension and gradually becoming fragmented into individualized forms of communication that are difficult to scrutinize publicly. In such a context, algorithmic personalization of political communication may progressively erode the shared public sphere, as different groups of citizens become exposed to divergent narratives of political reality without necessarily being aware of the differences between the information they receive. By integrating into digital infrastructures already characterized by recommendation systems and automated profiling, artificial intelligence tends to amplify dynamics that are already present within social media platforms, reinforcing processes such as personalized information selection, ideological polarization, and the formation of closed communication environments. Through the analysis of large volumes of behavioral data, algorithms can identify the content most likely to generate attention and emotional engagement, thereby favoring the dissemination of information that confirms existing beliefs rather than encouraging balanced deliberation among competing viewpoints. This mechanism contributes to the creation of so-called “filter bubbles”, within which individuals are exposed predominantly to opinions consistent with their own, reducing opportunities for dialogue with alternative perspectives and weakening the deliberative foundations of democracy (Pariser, 2011; Sunstein, 2018). Furthermore, the emergence of generative AI introduces an additional layer of complexity, as it enables the rapid production of highly persuasive and personalized content on an unprecedented scale. The ability to generate texts, images, audio recordings, and videos tailored to the psychological and social characteristics of specific segments of the electorate significantly increases the strategic potential of political campaigns. At the same time, however, it becomes increasingly difficult to distinguish between legitimate political communication, persuasive campaigning, and intentional manipulation. The principal concern therefore extends beyond the dissemination of false information to encompass the gradual loss of transparency in the processes through which individuals form their political opinions. For these reasons, the use of artificial intelligence in electoral processes requires a regulatory framework capable of balancing technological innovation with the protection of democratic principles. Establishing rules concerning algorithmic transparency, the use of personal data, the identification of AI-generated content, and the accountability of political actors represents a fundamental step toward limiting potential abuses. At the same time, promoting higher levels of digital literacy among citizens is essential so that individuals can develop the critical skills necessary to recognize manipulative content, understand how algorithmic systems operate, and participate consciously in the emerging digital public sphere. From this perspective, safeguarding democracy in the age of artificial intelligence cannot rely exclusively on technological or regulatory solutions. Rather, it requires an integrated approach involving public institutions, digital platforms, the scientific community, and civil society. Only through a combination of responsible regulation, technological transparency, and informed citizen awareness will it be possible to harness the potential of AI as a tool for political participation while preventing it from becoming a source of fragmentation, manipulation, and the erosion of democratic debate.
6. Conclusion
Throughout history, every technological transformation has produced significant changes in the ways in which societies organize communication, participation, and the dissemination of information. The digital revolution has further accelerated this process, profoundly altering the relationship between individuals, institutions, and communication tools. Within this scenario, politics represents one of the fields most strongly influenced by technological evolution, as the ability to communicate and build consensus has always been a central element of democratic competition (Gallo et al., 2022). The spread of online platforms and tools based on automated data processing has created new opportunities for citizen engagement, while also raising questions regarding information control, transparency, and the quality of public debate. In particular, the recent emergence of artificial intelligence has marked a new phase in this process, enabling the production and management of political content through previously unimaginable methods. To fully understand the consequences of this evolution, it is therefore necessary to examine not only the potential offered by new technologies but also the risks they may generate within contemporary democratic systems. On these foundations, the analysis of the role of artificial intelligence in electoral processes and in the new forms of political communication is developed. The introduction of artificial intelligence into electoral processes represents one of the most significant transformations in contemporary political communication, as it profoundly reshapes not only the ways in which candidates interact with voters but also the mechanisms through which citizens acquire information, form opinions, and participate in democratic life. The analysis conducted in this study demonstrates that AI should not be understood as an autonomous phenomenon that automatically generates either positive or negative consequences. Rather, it is a tool whose impact depends on the social, cultural, and institutional conditions within which it is developed and employed. The evolution of digital platforms and the rise of algorithmic systems have significantly expanded the possibilities for the personalization of political communication, enabling political actors to reach increasingly specific segments of the electorate through messages tailored based on behavioral data and individual preferences. On the one hand, this transformation creates new opportunities for civic participation and engagement by making political communication more responsive to citizens’ needs and fostering innovative forms of interaction between institutions and society. On the other hand, however, the growing capacity of artificial intelligence to generate, distribute, and optimize political content introduces new challenges concerning the transparency of communication processes, the protection of personal data, and the potential manipulation of public perceptions of reality. A comparison of different electoral experiences demonstrates that the effects of artificial intelligence are closely linked to the capacity of democratic systems to develop appropriate mechanisms of adaptation and response. The cases examined show that the presence of advanced technologies does not necessarily weaken democracy; rather, it can produce very different outcomes depending on the strength of the social fabric and the level of citizens’ trust in public institutions. The experience from Taiwan region illustrates how a society characterized by high levels of digital literacy, active civic networks, and effective information verification mechanisms can transform a potential informational threat into an opportunity for collective participation and the strengthening of democratic resilience. By contrast, the other nations have demonstrated how, in a context marked by deep political polarization and declining institutional trust, algorithmic systems can amplify pre-existing divisions, contributing to the fragmentation of the public sphere and the dissemination of manipulative narratives. From this perspective, artificial intelligence should be understood neither as an inherently destabilizing force for democratic systems nor as an automatic solution to the limitations of traditional political communication. Instead, AI functions as a multiplier of dynamics already present within the digital environment. It can strengthen participatory and inclusive processes when embedded within an ecosystem characterized by transparency, accountability, and critical awareness, but it can also intensify disinformation, polarization, and manipulation when deployed without adequate oversight and governance mechanisms (Battista & Gallina, 2025). Another central finding emerging from this analysis concerns the progressive transformation of political communication into an increasingly personalized practice. Through systems of profiling, political microtargeting, and automated content generation, contemporary electoral campaigns are moving beyond the traditional model of uniform mass communication, replacing it with multiple messages specifically adapted to the characteristics of different groups of voters. Although this evolution enhances communicative effectiveness, it also risks undermining the construction of a shared public sphere by fragmenting citizens’ informational experiences and making democratic deliberation across different perspectives increasingly difficult. For these reasons, the governance of artificial intelligence in electoral processes requires a multidimensional approach in which technological regulation constitutes only one component of a broader strategy. Alongside the establishment of rules concerning algorithmic transparency, the use of personal data, and the identification of AI-generated content, it is equally essential to invest in citizens’ digital literacy and to foster a critical culture of online engagement. The ability of future democracies to address the challenges posed by AI will depend not only on the effectiveness of available technological tools but also on their capacity to cultivate informed communities and institutions capable of governing these innovations responsibly. In conclusion, artificial intelligence represents a new dimension of political competition that is destined to exert an increasingly profound influence on participation, political communication, and the formation of public consensus. The principal challenge, therefore, is not to restrict technological progress but to ensure that it is directed toward objectives compatible with democratic principles. The future of digital political communication will depend on the ability to reconcile innovation with responsibility, ensuring that AI-based technologies strengthen rather than undermine the relationship between citizens and democratic institutions, and reinforce rather than erode the public trust upon which contemporary democratic systems ultimately depend.