Towards a conversational public administration? Public services, chatbots, and new organisational challenges for local administrationsABSTRACT
Introduction: In a context where public administrations are increasingly required to rethink their organizational and relational models in light of digital transformation, artificial intelligence (AI) emerges as a promising lever for innovation in local public services. In particular, the intersection between AI and institutional communication opens up new scenarios for the interaction between administrations and citizens, raising critical questions about efficiency, accessibility, and trust.
Objective: This study explores the organizational challenges and communication practices associated with the introduction of AI in Italian local governments, with a specific focus on the deployment of chatbots and the use of AI tools by public employees.
Methodology: The research adopts a qualitative methodology, based on 26 in-depth interviews with managers, officials, and heads of ICT and communication departments in public entities across three Italian regions (Lombardy, Emilia-Romagna, and Lazio), selected based on their level of digital maturity (Regional DESI Index, Politecnico di Milano, 2022).
Results: The findings reveal a heterogeneous and still experimental landscape: while there are promising opportunities in terms of automation and service improvement, structural weaknesses persist, particularly concerning organizational fragmentation, data governance, and the shortage of internal competencies. The use of chatbots in communication processes is widely viewed as potentially beneficial, yet its effectiveness is heavily dependent on the quality of informational infrastructures and the ability of AI to support trust-based relationships.
Conclusion: The study offers a reflection on the conditions necessary for the effective integration of AI into the broader eGovernment framework.
KEYWORDS:
Public sector; Public communication; Artificial intelligence; eGovernment
RIASSUNTO
Introduzione: in un contesto in cui le pubbliche amministrazioni sono sempre più chiamate a ripensare i propri modelli organizzativi e relazionali alla luce della trasformazione digitale, l’intelligenza artificiale (IA) emerge come una leva promettente per l’innovazione nei servizi pubblici locali. In particolare, l’intersezione tra IA e comunicazione istituzionale apre nuovi scenari per l’interazione tra amministrazioni e cittadini, sollevando questioni critiche riguardo a efficienza, accessibilità e fiducia.
Obiettivo: questo studio esplora le sfide organizzative e le pratiche comunicative associate all’introduzione dell’IA nelle amministrazioni locali italiane, con un focus specifico sull’impiego di chatbot e sull’uso di strumenti di IA da parte dei dipendenti pubblici.
Metodologia: la ricerca adotta una metodologia qualitativa, basata su 26 interviste in profondità con dirigenti, funzionari e responsabili dei dipartimenti ICT e di comunicazione in enti pubblici di tre regioni italiane (Lombardia, Emilia Romagna e Lazio), selezionati in base al loro livello di maturità digitale (Indice DESI Regionale, Politecnico di Milano, 2022).
Risultati: i risultati rivelano un panorama eterogeneo e ancora sperimentale: sebbene vi siano opportunità promettenti in termini di automazione e miglioramento dei servizi, persistono debolezze strutturali, in particolare riguardo alla frammentazione organizzativa, alla governance dei dati e alla carenza di competenze interne. L’uso dei chatbot nei processi comunicativi è ampiamente percepito come potenzialmente vantaggioso, ma la sua efficacia dipende in larga misura dalla qualità delle infrastrutture informative e dalla capacità dell’IA di sostenere relazioni basate sulla fiducia.
Conclusione: lo studio offre una riflessione sulle condizioni necessarie per l’efficace integrazione dell’IA all’interno del più ampio quadro dell’eGovernment.
PAROLE CHIAVE:
Pubblica aministrazione; Comunicazione pubblica; Intelligenza artificiale; eGovernment
1 INTRODUCTION
In recent years, the European Union has placed increasing emphasis on the development and regulation of artificial intelligence (AI), recognizing it as a strategic component for public sector innovation. This focus has led to the creation of a comprehensive regulatory and programmatic framework aimed at guiding the integration of AI in a responsible, transparent, and effective manner. The EU AI Act (Regulation (EU) 2024/1689) represents the cornerstone of the regulatory landscape for governing AI adoption across Europe, alongside strategic documents and operational tools issued by individual member states. The European Union is thus pursuing a dual objective: on the one hand, to contain and regulate the commercial push from private actors promoting AI-based solutions, and on the other, to foster their timely and informed adoption by public sector organizations. This approach seeks to promote innovation in public decision-making processes and in the services provided to citizens, while maintaining a cautious regulatory stance.
This study is situated within this broader context, with a specific focus on the Italian case and the organizational challenges currently faced by local administrations in adopting AI-based technologies.
While the digitalization efforts of the past two decades have enabled public administrations to equip themselves with infrastructures and technological tools aimed at significantly reshaping the public encounter (Goodsell, 1981) with citizens, current attention has shifted towards the adoption of systems capable of automating processes, providing virtual assistance, and supporting decision-making. The transition from analogue to digital has not merely involved the integration of new technologies, but has required a profound redefinition of internal organizational structures. In most cases, this has led administrations to rethink their workflows from a digitally native perspective, generating not only opportunities but also numerous challenges. Principles such as “digital first” and “once only”, as defined in the European digital strategy, exemplify how the public sector has been called upon to redesign administrative processes to align with digitalization logics, through the dematerialization of documents and time-saving practices that have progressively moved the interaction between citizens and public administration into fully online environments. Throughout this transition, governments have invested substantial financial and human resources in building institutional architectures capable of sustaining increasingly effective and efficient models of digital governance. In parallel, targeted interventions have been launched to boost the population’s digital literacy, with the aim of promoting widespread adoption and use of digital public services.
Within this context, a large majority (84%) of European citizens (an increase of 1% compared to 2024) believe that digital technologies will have a significant impact on access to online services (European Commission, 2025). Furthermore, 62% of Europeans view the introduction of robots and artificial intelligence in the workplace positively (ibid.), although this support is coupled with the recognition of the need for appropriate governance: 84% consider proper management of AI essential. These figures, however, reveal significant variation across European countries (ibid.): in Northern Europe, countries such as Denmark (86%), Finland (80%), Sweden (79%), and the Netherlands (77%) show a greater openness and trust in the adoption of AI and robotics in the workplace. This contrasts sharply with countries like Portugal (48%) and Greece (48%), where confidence in the positive impact of these technologies is markedly lower. Italy, with 61%, aligns closely with the European average (62%), trailing by just one percentage point. The adoption of artificial intelligence in the European public sector thus lies at the intersection of pronounced territorial asymmetries, discontinuities in innovation policies, and a growing push by decision-makers and policy-makers to embrace technologically advanced solutions.
Alongside the growing deployment of technological infrastructures, the platformization of public communication (Poell et al., 2019) has also contributed to the redefinition of digital relational spaces and prompted public administrations to simplify their traditionally technical language. This simplification extends beyond the mere rethinking of information transmission; it also contributes to the symbolic (re)construction of public space, making it recognizable both within and beyond the administrative environment (Lefebvre, 1991). Institutional websites, messaging apps, and official social media channels have become integrated with traditional public arenas -such as town squares, institutional campaigns, posters, and political rallies - fostering a convergence of languages, formats, and modes of content consumption (Jenkins, 2006). Within this evolving landscape of institutional public communication, language, design, and formats acquire a new role in representing institutions within a digital ecosystem increasingly characterized by narrative complexity and informational overload (Canel et al., 2018; Luoma-aho et al., 2021; Comunello et al., 2021; Lovari & Ducci, 2022).
1.1 Italian digitalization in the European context
The digitalization of public administration in Italy has evolved along a discontinuous trajectory, shaped to a significant extent by external drivers, most notably European directives beginning with the eEurope program (European Council of Feira, 2000). The first Italian eGovernment Plan, launched in 2002, sought to reorient administrative action toward a more horizontal and service-oriented structure. From the outset, however, the model faced critical challenges stemming from the institutional fragmentation that characterizes the Italian governance system. Administrative decentralization, together with the significant autonomy granted to local authorities, particularly the Regions, in matters of digitalization, produced marked heterogeneity in the implementation of digital strategies. This resulted in uneven outcomes in terms of innovation, accessibility, and systemic coherence.
Over the past two decades, several important regulatory and strategic instruments have been introduced, including the Digital Administration Code (2005), the Three-Year Plan for IT in Public Administration, and more recently the Italia Digitale 2026 initiative. Nevertheless, significant structural weaknesses remain. The European Commission’s Digital Decade Report (2025) confirms that the slow diffusion of digital skills among the population, and especially within the human capital of public administration, continues to represent one of the main barriers to the full realization of the digital paradigm. A further difficulty lies in moving beyond superficial technological adaptation. Italy has frequently attempted to transpose analog administrative processes into digital environments without undertaking genuine reengineering efforts, resulting in duplicated tasks rather than streamlined operations. This adaptive approach has preserved many of the rigidities typical of the traditional administrative model, ultimately revealing a limited inclination toward systemic transformation.
During the Covid-19 lockdown, alongside the urgent need to strengthen digital infrastructures, increasing attention was directed to the role of public communication within Italian administrations (Lovari, 2020; Lovari, D’Ambrosi, Bowen, 2020; D’Ambrosi et al., 2021). Press offices, public relations units, political figures, and institutional representatives gained centrality on digital platforms, which had until then been used by institutions in an exploratory and largely experimental manner. Public communication shifted from being an ancillary function to becoming a strategic lever for managing the relationship between institutions and citizens (Lovari & Ducci, 2022). In a context characterized by high uncertainty, limited mobility, and widespread reliance on digital services, the capacity of public administrations to communicate in a timely, clear, and accessible way proved essential not only for strengthening public trust but also, and above all, for ensuring access to essential services.
This trajectory has guided institutional public communication through several phases (ibid.), in parallel with the gradual development of digital government over the past two decades. Since the entry into force of Law 150/2000, which established the regulatory framework for public information and communication activities in Italy, the institutional communication ecosystem has undergone a profound transformation. Over time, there has been a significant evolution in the platforms employed, ranging from institutional websites to social media channels and, more recently, to the adoption of AI-based tools such as chatbots, virtual assistants, and automated messaging applications. These developments reflect the administrations’ need to maintain a presence in an ever-changing digital environment and to reach citizens through the channels and modes of communication that are embedded in their everyday lives.
The growing strategic relevance of digital communication is further confirmed by the recent Italian Public Administration Decree (D.L. no. 25/2025), which has formally introduced the role of the social media manager within public institutions. Within the framework of the National Recovery and Resilience Plan (PNRR), Italy has reinforced its strategic orientation toward digital government. Initiatives such as Mission M1C1 “Digitalization, Innovation, and Security in the Public Administration” and, in particular, component 1.4.1 focused on “Citizen Experience in Public Services” clearly demonstrate the intention to reposition the citizen at the center of administrative action. Finally, increasing attention has been directed to the integration of artificial intelligence solutions, culminating in the adoption of specific guidelines for the use of AI within public administrations.
2 THEORETICAL FRAMEWORK
The literature on eGovernment, initially focused primarily on infrastructural development, has progressively shifted toward a relational dimension of digital governance (Davis, 1989; Carter et al., 2005; Banniester et al., 2014; Luoma-aho, 2021; Solito & Materassi, 2021; Bessières, 2021; Lovari & Ducci, 2022). Scholarly attention has moved beyond the analysis of technological deployment, extending to the examination of the impact and the generation of public value (Deidda Gagliardo, 2015) that digital technologies can create (van Noordt & Misuraca, 2022; 2023). It is not by chance that the first adoptions of AI by public organizations frequently involve the deployment of virtual assistants based on generative AI technologies. This tendency reflects the institutional and administrative drive to identify digital solutions capable of enhancing the relationship with citizens rapidly and at relatively low cost, although often with limited quality and effectiveness.
The literature on public communication (Macnamara, 2018; Grunig, 2000; Lovari & Bowen, 2020) has increasingly emphasized the central role of transparency, listening, and participation in fostering trust-based relationships between citizens and institutions. The relational paradigm, in contrast to unidirectional and persuasive communication models, underscores the need to move beyond top-down approaches toward more open, dialogical, and co-creative practices. Within this perspective, trust is not simply a desirable outcome but serves as an indicator of the democratic quality of communicative processes (Smillie & Scharfbillig, 2024).
Within this debate, the introduction of artificial intelligence into the communicative processes of public administrations raises new theoretical and practical challenges. On the one hand, AI-based technologies such as chatbots, virtual assistants, and automated response systems promise efficiency, accessibility, and continuity in the delivery of public information services. On the other hand, they generate significant relational concerns: what are the effects of replacing human interaction with algorithmic interfaces on citizens’ perceptions of trust? How does the nature of institutional communication change when the interlocutor is no longer a person but a machine? How does this shift affect perceptions of availability, transparency, and responsiveness on the part of the public sector?
As Gerlich (2024) argues, chatbots should be situated within the broader category of “virtual humans” (VHs), whose increasing anthropomorphization blurs the boundaries between human and machine. The attribution of human-like features to non-human entities raises complex ethical and relational dilemmas, especially concerning the authenticity of interactions and the conditions under which citizens place their trust in digital interfaces (ibidem). Issues of transparency about the virtual nature of these systems, the exploitation of user data, and the biases embedded in the large language models (LLMs) that underpin VHs further complicate their institutional use (ibidem).Similarly, Opromolla et al. (2024) conceptualize artificial intelligence, of which chatbots are one application, as systems endowed with human-like capacities for learning and decision-making.
Their analysis emphasizes the importance of signalling when AI is activated, what it is doing, which data it is using, and for what purposes. This call for transparency resonates strongly with debates on accountability in digital governance and underscores the need to avoid the opacity that often characterizes algorithmic processes. These perspectives suggest that the introduction of AI in public communication cannot be understood solely as a matter of technological efficiency. Rather, it entails a reconfiguration of symbolic and relational dynamics in citizen-state interactions. Anthropomorphization, opacity, and data exploitation not only affect trust but also risk reshaping the very foundations of democratic communication, where principles of openness, accountability, and inclusiveness remain central.
To address these questions, this study adopts the theoretical framework developed by Lovari et al. (2021), which focuses on the dialectic between openness and closeness in public communication and identifies ethics as a key driver of trust. This framework, structured around three dimensions-transparency, participation, and collaboration-enables the analysis of public communication along a continuum ranging from trust to distrust and from proximity to symbolic distance. The contribution of this study builds on and extends this model by incorporating artificial intelligence as a systemic variable. Specifically, it explores how the use of automated tools in digital public services may reshape not only institutional interaction but also the perceived structure of communicative openness.
The study pursues two research questions: (RQ1) What are the current organizational challenges that Italian local municipalities face in adopting AI-based systems?. (RQ2) What role is artificial intelligence playing in shaping communicative practices that enable dialogue between local governments and citizens?
By addressing these questions, the research seeks to contribute to the eGovernment literature by providing a more nuanced and critical understanding of digital government in the age of artificial intelligence.
3 METHODOLOGY
To address the research questions, a series of in-depth interviews were carried out between 2024 and 2025 with directors, officials, and ICT and communication managers from Italian local authorities (Chart 1). The participants’ accounts centered on digital public services, institutional communication, and the current as well as future role of artificial intelligence. The methodological choice to involve key stakeholders made it possible to examine in depth the role of institutional actors in digital transition processes, thereby providing a nuanced understanding of existing challenges and potential trajectories for future development.
This study adopts a qualitative approach based on 26 in-depth interviews, each with an average duration of approximately one hour. The selection of participants, determined by availability, was conducted through a purposive and strategic sampling method informed by the most recent data from the Regional DESI Index (Politecnico di Milano, 2022). According to these data, the regions of Lazio, Lombardy, and Emilia-Romagna represent the territories with the highest levels of digital maturity in Italy. Digital maturity levels are calculated on the basis of the digital skills of human capital, connectivity, the availability of digital services, and the introduction of digital technologies within public institutions. The decision to focus on these regions reflects the aim of examining administrative contexts that have initiated or consolidated digital transformation processes, with the objective of identifying both emerging practices and structural challenges.
The data collection extended beyond regional administrations to include provincial capitals within the selected territories. This methodological choice enabled a more in-depth exploration of the local level, where many of the most frequently used public services-such as civil registry, urban planning, and social services-are concentrated, and where the practical implementation of regional digitalization policies takes place. In this respect, the study examined the relationship between multi-level governance, organizational structures, and the perception of digital transformation among local public actors.
The interview guide was structured into three main thematic sections: (1) the internal organization of the public body and its processes; (2) communication practices and the relationship with citizens, with particular attention to the construction of trust; and (3) the adoption and use of artificial intelligence technologies in communication processes and public service delivery. This structure made it possible to investigate the intersections between public communication, digital transformation, and the use of AI, highlighting how these dimensions interact and influence one another in the innovation processes of local administrations. All interviews were fully transcribed and analyzed thematically in order to identify recurring patterns, interpretive divergences, and shared strategic orientations among the institutional actors involved.
4 RESULTS AND DISCUSSION
The sample selected for this qualitative study consists of 26 interviewees from three Italian regions: Lombardy, Emilia-Romagna, and Lazio. These regions were chosen on the basis of their levels of digital maturity as indicated by the 2022 Regional DESI Index. The territorial composition of the sample was designed to include representatives from both regional administrations and provincial capitals, thereby ensuring a multilevel perspective on digital transformation and on the evolving relationship between citizens and public administrations.
Within the sample, Lombardy is the most represented region, with 14 interviewees. Of these, 2 are employed within the regional administration, while the remaining 12 are affiliated with provincial capital municipalities, including Milano, Como, Lecco, Lodi, Cremona, Mantova, Pavia, and Varese. The predominance of Lombardy reflects not only its demographic and economic weight but also its consolidated level of digital maturity. Emilia-Romagna follows with 8 interviewees, 5 from the regional administration and 3 from provincial municipalities (Ferrara, Rimini, and Reggio Emilia), which highlights the strength of regional digital governance and its operational articulation at the local level. Lazio is represented by 4 participants, including 2 in managerial positions and 2 municipal officers from the cities of Rome and Rieti. The inclusion of two interviewees from the City of Rome is particularly significant, given the dual role of the capital as both a local government and a central institutional hub, with strategic importance in Italy’s intergovernmental system.
With respect to gender distribution, the sample is nearly balanced, comprising 17 men and 9 women. While gender was not a central analytical variable, this distribution allows for the observation of possible gender-related differences in representation. Regarding institutional roles, participants hold senior or highly qualified positions: 4 Directors, 7 Senior Executive Officers (Dirigenti), 5 Managers (Responsabili), and 10 Highly Qualified Staff (Elevata qualificazione). This composition is consistent with the study’s objective of gathering well-informed and strategic perspectives on the organizational, communicative, and technological processes associated with digital transformation and with the adoption of artificial intelligence in local government contexts.
4.1 The internal organization for the digitalization of local administrations
The analysis of the interviews highlights substantial heterogeneity in the internal configuration of the administrations involved, which makes direct comparison between entities particularly complex. Although the 26 interviewees all operate in areas related to digitalization, information services, communication, or the management of the digital agenda, they hold positions with different titles, responsibilities, and hierarchical levels. The administrations present non-standardized organizational structures that are often shaped by institutional size, the availability of internal resources, and the presence or absence of strategic digital governance.
In some municipalities, responsibility for digital innovation coincides with the role of IT directors or area managers, whereas in other cases it is entrusted to civil servants with high-level qualifications but without senior managerial responsibilities. Particularly in small and medium-sized municipalities, the presence of intermediate figures is more common, resulting in a fragmented distribution of duties across ICT, citizen services, communication, and digital transformation. Only in a few cases, most notably in large metropolitan cities such as Milan and Rome, can more structured organizational settings be observed, with dedicated operational units and formalized roles.
While regional administrations tend to be better equipped in organizational terms, they also reveal differences in the allocation of responsibilities among general directorates, sector managers, and digital agenda coordinators. Moreover, the coexistence of administrative, technical, and communication functions, and their variable positioning within organizational charts, contributes to a fragmented scenario that underscores the absence of a nationally shared model for the governance of technological innovation. This structural variability leads to uneven decision-making processes and weak coordination across sectors, which ultimately undermines the systematic and coherent adoption of advanced technological solutions such as artificial intelligence.
On the basis of an analytical synthesis of the profiles, the interviewees were grouped into five macro professional categories: (1) heads of information services, (2) area directors with expertise in ICT and service management, (3) highly qualified officers in public services, (4) digital agenda coordinators, and (5) institutional communication and press office managers. This classification reflects the organizational diversity of the public entities involved and the multiplicity of approaches to managing digital innovation processes.
4.2 Artificial Intelligence in the Organizational Practices and Public Services of Local Governments
The thematic analysis of the interviews highlights both the opportunities perceived (Fig. 1) and the challenges reported (Fig. 2) by respondents regarding the adoption of artificial intelligence within local governments. The opportunities identified can be grouped into four main areas: the management of digital data through functional databases, the automation of low-complexity tasks, the analysis of processes and service usage patterns, and time savings in responding to citizen requests.
In particular, the introduction of automated systems is regarded as a strategic lever for enhancing operational efficiency, especially in repetitive and standardized administrative processes. Such systems allow for the rationalization of internal resources and a reduction in staff workload, particularly in domains such as civil registry services, document management, and appointment scheduling, which are frequently characterized by routine tasks. Furthermore, the capacity of intelligent systems to analyze data flows and provide real-time insights is seen as having significant transformative potential for both the design of public services and the continuous monitoring of citizen satisfaction and administrative performance.
However, this potential is constrained by a set of systemic issues that hinder the full integration of AI within the public sector. One of the main barriers identified concerns data interoperability, that is, the still severely limited capacity of administrative databases to communicate in an integrated and efficient manner. This limitation is compounded by the challenge of ensuring real-time database updates, which are essential for the effective functioning of predictive algorithms and informational chatbots.
The development of technical skills represents another critical issue (European Commission, 2025). Many interviewees reported a shortage of adequately trained internal staff, pointing to a structural dependence on external providers. This reliance frequently results in a near-systematic outsourcing dynamic, which in turn generates difficulties in managing public-private contracts, particularly with regard to costs, the reliability of supplies, and the suitability of technological tools in relation to actual organizational needs.
However, the acquisition of digital skills in the Italian public sector often depends on the individual initiative of civil servants, who frequently resort to self-learning through technical manuals and software programs to address the operational needs of their offices. Although central initiatives such as the Syllabus for the Public Administration have been introduced, their effective adoption and practical use remain difficult to evaluate. This situation suggests that regional policies and organizational frameworks have not yet succeeded in establishing coherent and systematic pathways for digital skills development, thereby leaving much of the responsibility in the hands of individual professionals. Such dynamics highlight the uneven distribution of competences across administrations and the persistence of structural weaknesses in the institutional capacity to promote digital literacy as a collective and strategic resource.
The combination of these elements depicts a complex scenario in which the transformative potential of artificial intelligence is widely acknowledged, yet its implementation remains dependent on actual organizational capacity, the maturity of data governance, and structural policies for skills development. Artificial intelligence therefore emerges simultaneously as a catalyst for innovation and as a source of strain for administrative systems that are not yet fully adapted to the dynamics of digital transformation.
4.3 Is a chatbot for dialogue between administrations and citizens possible?
Within the broader framework of digital transformation in public administration, the introduction of artificial intelligence into communication processes is increasingly emerging as a strategic frontier. The interviews conducted reveal broad consensus on the value of equipping local administrations with automated tools for public interaction, such as chatbots. All interviewees recognized the potential of these technologies to enhance the efficiency of institutional communication, particularly in areas characterized by high volumes of standardized information requests, including Public Relations Offices and tourism information services.
At the same time, this general agreement on the usefulness of chatbots is tempered by a cautious and critical perspective regarding the conditions necessary for their effective and responsible implementation. One of the most frequently mentioned concerns relates to the structural challenges of public data management, which is often fragmented, outdated, and poorly interoperable. These shortcomings represent a significant technical barrier to the development of chatbot systems that can be considered genuinely intelligent.
The variety of citizens’ queries requires high-quality and well-structured public data. (Interviewee 19)
Technical limitations, such as the ability to understand the diverse range of queries citizens use when interacting with chatbots, illustrate how the effectiveness of automated interaction is highly dependent on the underlying technology, which often fails to adapt to the dynamics of unmediated social interaction (Thompson, 1995). In this respect, the performance of chatbots in the private sector was frequently cited as a benchmark that should not be underestimated. At the same time, it was regarded as a model that is difficult to replicate in the public sector without targeted investments, particularly when relying on third-party solutions that are often misaligned with the needs of the administration and, consequently, those of citizens.
We can talk about artificial intelligence, but we should actually be talking about the solutions being proposed to us. They’re already outdated. (Interviewee 26)
Although still limited, the ongoing experiments indicate a degree of institutional openness toward the adoption of AI-based solutions. This openness, however, is frequently constrained by a combination of technical and managerial limitations as well as organizational culture. A recent survey conducted by the Italian Digital Agency (Agenzia per l’Italia Digitale, 2025) confirms the growing presence of artificial intelligence in central administrations, documenting 120 projects aimed at enhancing operational efficiency, data management, and service access. At the same time, the survey identifies persistent structural weaknesses, including a high level of outsourcing (50 percent) and a significant shortage of in-house professional expertise. More than 60 percent of the projects involve the use of chatbots, yet the quality of training data remains a critical issue. These findings underscore the need for sustained investments in skills, data governance, and innovative procurement practices, concerns that were also echoed by the interviewees in this study.
Furthermore, the thematic analysis reveals that the integration of artificial intelligence into institutional communication is closely linked to the issue of public trust (Smillie & Scharfbillig, 2024). Several interviewees emphasized that trust is not determined by the tool itself but rather by how citizen requests are managed, particularly whether chatbots are able to provide responses based on real-time and updated data.
Trust often depends on the target audience, and everything revolves around how the request is handled. In the end, it all comes down to how precise, effective, and timely the response is. (Interviewee 10)
This underscores that the core challenge does not lie in replacing human interaction but rather in the capacity of AI to achieve equivalent performance in terms of clarity, personalization, timeliness, and the security of data sharing.
Overall, the role of artificial intelligence in the communicative practices of local administrations is perceived as potentially transformative, yet constrained by technical, organizational, and cultural factors. When appropriately designed, chatbots can provide a useful interface for institutional dialogue. However, in order for them to genuinely enhance the relationship between citizens and public administrations, it is crucial to address critical issues related to data quality, decision-making processes, and design standards. Trust therefore remains the central element that must be preserved and cultivated, even in the adoption of the most advanced technologies.
6 CONCLUSION
The findings of this study provide a critical interpretation of the ongoing transformations within Italian local administrations regarding the introduction of artificial intelligence systems. In relation to the first research question (RQ1), the results clearly demonstrate that organizational challenges are numerous and multi-layered, spanning structural, cultural, and technological dimensions. The main obstacles include internal organizational fragmentation, the absence of a coherent and shared data governance framework, and the difficulty of establishing professional roles specifically designed to foster innovation, particularly those related to AI management and data quality. Moreover, the predominance of procurement models that are not tailored to innovation, combined with a strong reliance on external providers, continues to limit the strategic autonomy of public bodies.
With respect to the second research question (RQ2), the study highlights a growing interest in artificial intelligence as a tool for supporting communication and relationship-building practices between administrations and citizens. Chatbots emerge as the most widely explored application, particularly within public information offices and tourism-related services. While respondents generally recognized the potential of these tools to enhance efficiency, accessibility, and timeliness, doubts persist about their actual ability to replace, even partially, human interaction, especially in contexts where trust plays a key role. Communication continues to be regarded as a strategic lever for building institutional relationships grounded in listening, transparency, and personalized interaction.
Nonetheless, the adoption of AI within local administrations remains embryonic and uneven, often motivated more by the desire to follow technological trends than by a systemic transformation of citizen engagement models. Technologies are frequently introduced in experimental phases, without being fully integrated into the organizational and communicative frameworks of public administrations. The study therefore underscores the importance of approaching AI adoption not merely as a technical innovation but as a cultural and organizational process, one that requires the development of new competencies, governance models, and relational paradigms that are more open, inclusive, and oriented toward the generation of public value(Moore, 1995; Deidda Gagliardo, 2015; Van Noordt & Misuraca, 2020; Noordt et al., 2023).
In this sense, the findings point to the central tension between algorithmic efficiency and democratic trust in public services. While the promise of automation lies in streamlining procedures and reducing administrative burdens, its democratic legitimacy depends on the capacity of institutions to cultivate trust, transparency, and accountability in their interactions with citizens(Canel & Luoma-aho, 2018; Lovari et al., 2020; Lovari & Ducci, 2022). The theory of public value(Moore, 1995) provides a useful lens to interpret this tension, as it highlights that the effectiveness of digital transformation cannot be measured solely in terms of efficiency gains but must also consider the broader contribution of technology to the public good, collective legitimacy, and democratic participation.
The theoretical implications of this study lie precisely in this dialectic: AI-driven tools should not be examined only as instruments of organizational optimization but as socio-technical arrangements that reshape the symbolic, relational, and normative foundations of public communication. Future research should therefore investigate how algorithmic mediation redefines the conditions of trust, the quality of democratic dialogue, and the capacity of digital services to generate genuine public value. By doing so, the analysis would move beyond a descriptive approach and contribute to a more original and innovative understanding of AI in the context of digital governance.
Acknowledgments:
Not applicable.
References
-
AGENZIA PER L’ITALIA DIGITALE. L’intelligenza artificiale nella pubblica amministrazione: ricognizione delle PA centrali. Rapporto 2025. Disponibile in: https://www.agid.gov.it/sites/agid/files/2025-06/ AGID_Ricognizione_Progetti_AI_Rapporto_2025_V_2_2_10giu25.pdf Accesso il: 2 set. 2025.
» https://www.agid.gov.it/sites/agid/files/2025-06/ AGID_Ricognizione_Progetti_AI_Rapporto_2025_V_2_2_10giu25.pdf -
BANNISTER, F.; CONNOLLY, R. ICT, public values and transformative government: A framework and programme for research. Government Information Quarterly, London, v. 31, n. 1, p. 119-128, 2014. DOI: https://doi.org/10.1016/j.giq.2013.06.002
» https://doi.org/10.1016/j.giq.2013.06.002 - BESSIÈRES, D. The hybridity of public communication: on old component still a sign of modernity in France. Sociologia della Comunicazione, Milan, v. 61, p. 36-51, 2021.
- CANEL, M. J.; LUOMA-AHO, V. Public sector communication: closing gaps between citizens and public organizations. Chichester: Wiley-Blackwell, 2018.
-
CARTER, L.; BÉLANGER, F. The utilization of e-government services: Citizen trust, innovation, and acceptance factors. Information Systems Journal, Oxford, v. 15, n. 1, p. 5-25, 2005. DOI: https://doi.org/10.1111/j.1365-2575.2005.00183.x
» https://doi.org/10.1111/j.1365-2575.2005.00183.x - COMUNELLO, F.; MASSA, A.; IERACITANO, F.; MARINELLI, A. Public sector communication professions in the Twitter-sphere. Sociologia della Comunicazione, Milan, v. 61, p. 90-108, 2021.
- D’AMBROSI, L., INIESTA, I., PARITO, M., PÉREZ-CALLE, R. Fighting disinformation. The impact of the Covid-19 on youth trust in European Institutions. Sociologia della Comunicazione, Milan, v. 61, p. 71-89, 2021.
-
DAVIS, F. D. Perceived usefulness, perceived ease of use, and user acceptance of information technology. MIS Quarterly, Minneapolis, v. 13, n. 3, p. 319-339, 1989. DOI: https://doi.org/10.2307/249008
» https://doi.org/10.2307/249008 - DEIDDA GAGLIARDO, E. Il valore pubblico: la nuova frontiera delle performance. v. 30. Roma: RIREA, 2015.
-
EUROPEAN COMMISSION; DIRECTORATE-GENERAL FOR COMMUNICATIONS NETWORKS, CONTENT AND TECHNOLOGY; CAPGEMINI; IDC; POLITECNICO DI MILANO; SOGETI. EGovernment benchmark 2025 - building online, user-friendly, and interoperable services - factsheets. Luxembourg: Publications Office of the European Union, 2025. Disponibile in: https://data.europa.eu/doi/10.2759/9177128 Accesso il: 2 set. 2025.
» https://data.europa.eu/doi/10.2759/9177128 -
EUROPEAN COMMISSION. Digital Decade 2024: Europe’s digital transformation. 2024. Disponibile in: https://digital-strategy.ec.europa.eu/en/factpages/digital-decade-2024-report-country-fact-pages Accesso il: 2 set. 2025.
» https://digital-strategy.ec.europa.eu/en/factpages/digital-decade-2024-report-country-fact-pages -
EUROPEAN COMMISSION. Artificial intelligence and the future of work (Special Eurobarometer No. SP554). Directorate-General for Employment, Social Affairs and Inclusion, 2025. Disponibile in: https://europa.eu/eurobarometer/surveys/detail/3222 Accesso il: 2 set. 2025.
» https://europa.eu/eurobarometer/surveys/detail/3222 -
EUROPEAN COMMISSION. Science and technology (Special Eurobarometer No. SP557). Directorate-General for Research and Innovation, 2025. Disponibile in: https://europa.eu/eurobarometer/surveys/detail/3227 Accesso il: 2 set. 2025.
» https://europa.eu/eurobarometer/surveys/detail/3227 -
GERLICH, M. Societal perceptions and acceptance of virtual humans: trust and ethics across different contexts. Social Sciences, Basileia, v. 13, n. 10, p. 516, 2024. DOI: https://doi.org/10.3390/socsci13100516
» https://doi.org/10.3390/socsci13100516 - GOODSELL, C. T. The public encounter: where state and citizen meet. Bloomington: Indiana University Press, 1981.
- GRUNIG, J. E. Collectivism, collaboration, and societal corporatism as core professional values in public relations. Journal of Public Relations Research, Philadelphia, v. 12, n. 1, p. 23-48, 2000.
- KATTEL, R.; DRECHSLER, W.; KARO, E. How to make an entrepreneurial state: why innovation needs bureaucracy. New Haven: Yale University Press, 2022.
-
KUZIEMSKI, M.; MISURACA, G. AI governance in the public sector: three tales from the frontiers of automated decision-making in democratic settings. Telecommunications Policy, Oxford, v. 44, n. 6, art. 101976, 2020. DOI: https://doi.org/10.1016/j.telpol.2020.101976
» https://doi.org/10.1016/j.telpol.2020.101976 - JENKINS, H. Convergence culture: where old and new media collide. New York, NY: New York University Press, 2006.
- LEFEBVRE, H. The production of space. Oxford: Blackwell, 1991.
-
LINDGREN, I.; MADSEN, C. Ø.; HOFMANN, S.; MELIN, U. Close encounters of the digital kind: a research agenda for the digitalization of public services. Government Information Quarterly, London, v. 36, n. 3, art. 101377, 2019. DOI: https://doi.org/10.1016/j.giq.2019.03.002
» https://doi.org/10.1016/j.giq.2019.03.002 -
LOVARI, A.; BOWEN, S. A. Social media in disaster communication: a case study of strategies, barriers, and ethical implications. Journal of Public Affairs, Oxford, v. 19, n. 3, p. 1-9, 2019. DOI: https://doi.org/10.1002/pa.1967
» https://doi.org/10.1002/pa.1967 - LOVARI, A.; DUCCI, G. Comunicazione Pubblica: istituzioni, pratiche, piattaforme. Milano: Mondadori Università, 2022.
-
LOVARI, A.; D’AMBROSI, L.; BOWEN, S. A. Re-connecting voices. The (new) strategic role of public sector communication after the Covid-19 crisis. Partecipazione e Conflitto, Lecce, v. 13, n. 2, p. 970-989, 2020. DOI: https://doi.org/10.1285/i20356609v13i2p970
» https://doi.org/10.1285/i20356609v13i2p970 - LUOMA-AHO, V.; CANEL, M.-J.; BOWDEN, J.; EK, E.; VAINIÖMÄKI, V. Citizen engagement and the Covid-19 mask communication. Sociologia della Comunicazione, Milan, v. 61, p. 20-35, 2021.
- MACNAMARA, J. Public relations and post-communication: addressing a paradox in public communication. Public Relations Journal, New York, v. 11, n. 3, p. 1-20, 2018.
-
MIKALEF, P.; GUPTA, M. Artificial intelligence capability: conceptualization, measurement calibration, and empirical study on its impact on organizational creativity and firm performance. Information & Management, Amsterdã, v. 58, n. 1, art. 103434, 2021. DOI: https://doi.org/10.1016/j.im.2021.103434
» https://doi.org/10.1016/j.im.2021.103434 - OPROMOLLA, A.; PARISI, L.; DE MATTEIS, G.; PIZOLLI, R.; TRODINI, L. L’innovazione dei servizipubbliciattraversol’intelligenzaartificiale: scenaridiinnovazioneemergentitra i giovaniadulti. Mediascapes Journal, Roma, v. 23, n. 1, p. 27-45, 2024.
-
POELL, T.; NIEBORG, D.; VAN DIJCK, J. Concepts of the digital society: platformisation. Internet Policy Review, Berlim, v. 8, n. 4, p. 1-13, 2019. DOI: https://doi.org/10.14763/2019.4.1425
» https://doi.org/10.14763/2019.4.1425 - POLITECNICO DI MILANO - OSSERVATORI DIGITAL INNOVATION. Il posizionamento delle regioni e province autonome sul DESI regionale 2022. Osservatorio Agenda Digitale, 2023.
-
SMILLIE, L.; SCHARFBILLIG, M. Trustworthy Public Communications. Luxembourg: Publications Office of the European Union, 2024. DOI: https://doi.org/10.2760/695605
» https://doi.org/10.2760/695605 - SOLITO, L.; MATERASSI, L. If a picture is not worth a thousand words: digital infographics use during the Covid-19 pandemic crisis. Sociologia della Comunicazione, Milan, v. 61, p. 52-70, 2021.
- THOMPSON, J. B. The media and modernity: a social theory of the media. Cambridge: Polity Press; Stanford: Stanford University Press, 1995.
- TOLL, D.; LINDGREN, I.; MELIN, U.; MADSEN, C. Ø. Artificial intelligence in Swedish policies: values, benefits, considerations and risks. Linköping: Linköping University, Research Centre for Government IT, 2024.
-
VAN NOORDT, C. Public value creation with artificial intelligence technologies in public administration. TalTech Press, 2023. DOI: https://doi.org/10.23658/taltech.58/2023
» https://doi.org/10.23658/taltech.58/2023 -
VAN NOORDT, C.; MISURACA, G. Artificial intelligence for the public sector: results of landscaping the use of AI in government across the European Union. Government Information Quarterly, London, v. 39, n. 3, art. 101714, 2022. https://doi.org/10.1016/j.giq.2022.101714
» https://doi.org/10.1016/j.giq.2022.101714 -
WIRTZ, B. W.; WEYERER, J. C.; GEYER, C. Artificial intelligence and the public sector - applications and challenges. International Journal of Public Administration, New York, v. 42, n. 7, p. 596-615, 2019.https://doi.org/10.1080/01900692.2018.1498103
» https://doi.org/10.1080/01900692.2018.1498103
Edited by
-
Editor guest:
Maria Giovanna Guedes Farias https://orcid.org/0000-0002-2690-3350
Not applicable.
Fonte: [author’s elaboration based on data]
Fonte: [author’s elaboration based on data]
