Guest editorial: Generative AI and the government: looking through, thinking forward, contextualizing

The launch of ChatGPT, a prominent generative artificial intelligence (AI) application, sparked widespread interest in its potential uses. Scholars and professionals are exploring the scope of generative AI’s applications but, at the same time, there is growing concern about its unintended consequences.These concerns have led to serious discussions about the need for a global regulatory framework that addresses in a comprehensive and consistent way both the development and use of generative AI. Organizations like the OECD, EU and UN are leading a global effort to create a consensus on how generative AI should be developed and used. It would be a desirable goal to be able to define a regulatory framework that both promotes technology’s growth and ensures its use aligns with existing legal, ethical and societal norms.Generative AI refers to algorithms that can produce new content such as text, images and code. While it is a form of machine learning, its applications are widespread, from product design and business optimization to complex HR processes. This technology also has the potential to transform how we use big data in areas like governance and policymaking (Visvizi, 2022; Visvizi and Bodziany, 2021).Despite its seemingly limitless applications due to the vast neural networks it uses, generative AI still has significant room for improvement. Key issues that need to be addressed include bias, imprecision, faulty logic and concerns related to ethics and legal frameworks (Dwivedi et al., 2023). The role of government is crucial in regulating generative AI. This is a complex issue because the government must both guide the ethical development of technology for societal benefit (like fighting poverty or managing traffic) and be wary of its potential to be misused.In this context, this Special Issue aimed to explore the nexus between the government and generative AI to identify and explore impacts on consolidated modes of governance, policymaking, strategy design and stakeholders’ involvement.Government stands at the core of the necessary regulatory process geared toward making generative AI the source of goods and services and fostering economic and social development. The regulatory framework, therefore, can facilitate or hinder the development of this technology.From a different angle, in several ways, generative AI offers the potential of taking our thinking of the value of technology and data analysis in the domains of governance, policymaking, strategy design and stakeholders’ involvement. Moreover, it can not only be the source for relevant managerial and implications but can also stimulate the development of new entrepreneurial opportunities while acting at the same time as a source for innovation and sustainable-oriented practices (Perucica and Andjelkovic, 2022; Baiocco et al., 2024).Recent literature is seeking to explore the multi-sided potential of generative AI in government (Pandey, 2026). However, the disruptive phase is still in its early stages, and this implies the need for a deeper exploration of the effects of technological changes at various government levels, not only enveloping local and national contexts, but also exploring managerial, organizational and community level (Korzynski et al., 2023) according to both an individual and ecosystems perspective to analyze actors’ practices and behaviors.This Special Issue collected contributions that analyze the role of government in guiding generative AI deployment by detecting the limits, opportunities, regulatory frameworks, stakeholders deriving from this application. The goal, however, was not only to attract studies with a government perspective but to shed light on the heterogeneity of contexts that can host generative AI, including the business–government nexus, the civic society–government nexus and the politics–economic policy nexus. Although the regulatory challenge is a matter of urgent resolution, a correct application of AI in business life should necessarily be based on the enrichment of digital skills for knowledge creation and the emergence of innovation (Haefner et al., 2021), in the light of Industry 5.0 principles that require the integration of digital and human components (Troisi et al., 2022, 2024).Therefore, a full exploration of generative AI in government should encompass the legal, ethical, administrative and managerial perspectives, examined according to a human, social and cognitive multidisciplinary standpoint.The different articles included in this Special Issue investigate the different shades of AI application to different contexts, at different levels of analysis, by raising political, social, economic and managerial implications.As synthesized in Table 1, the peculiarities (research objectives, methodological architectures and scope) of each article contribute to provide an overview of the complex set of topics that the use of generative AI puts forward. Starting from the common willingness to detect barriers and drivers for the adoption of AI in public sector, some articles shed light on the issue of technical criticalities (transparency, security, accountability), others emphasize the need for a digital humanism and, hence, for the enrichment of knowledge and digital literacy, whereas others emphasize emotional and reputational risks or the societal implications.The article by Jitendra K. Pandey, titled “Unlocking the Power and Future Potential of Generative AI in Government Transformation,” places strategy work inside organizational realities, with a strong emphasis on contextual fit. Drawing on Indian public-sector cases and management surveys, analyzed through exploratory data analysis, the author shows how adoption patterns hinge on managerial capacity, process redesign, data stewardship and staff training, not only on algorithmic performance. Findings underline promise around responsiveness, decision support and cost containment, while acknowledging measurement gaps and change-management constraints that administrators frequently report. The paper calls for staged pilots, resource alignment and monitoring routines that watch quality, security and equity, rather than headline performance alone. An additional contribution lies in its explicit bridge between digital-government scholarship and practice, offering language that CIOs, auditors and line managers can use during procurement and evaluation.Generative AI can be fruitfully employed in media channels, since news carries authority during technological transitions. Yucong Lao and Yukun You, in “Unraveling Generative AI in BBC News: Application, impact, literacy and governance,” describe news stories about generative systems, then extract key elements of “AI literacy” that matter for audiences beyond classrooms, including risk awareness, claims evaluation and responsible use in everyday settings. Their analysis explicitly links literacy with governance debates, arguing that policy bodies and broadcasters share obligations that shape citizens’ understanding of automation, misinformation and institutional safeguards. The article maps a space where public communication and regulatory practice meet, with careful attention paid to definitional clarity, domain-specific pedagogy and administrative implications. Evidence comes from a wide corpus of BBC items, read methodically rather than anecdotally, which grounds the literacy components in everyday narratives that citizens already encounter.Education sector constitutes a fertile field of application of generative AI, by potentially modifying how students interpret and use language models that are currently largely adopted in teaching and assessment. Mario Testa, Maddalena Della Volpe, Antonio D’Amato and Adriana Apuzzo, in “Does Gender Impact the Relationship Between Perceived Value and Intentions of Use of Natural Language Processing Models?”, survey 562 students from the University of Salerno, then estimate models that test whether perceived value predicts intention, and whether gender moderates that relationship. Results show a strong positive association between perceived value and intention; interaction terms reveal a steeper slope for males within several disciplinary clusters, while humanities show smaller differences across gender groups. Policy suggestions include targeted training, emphasis on demonstrable educational use and programs that support confidence and capability among women who report lower self-assessed digital skills.Risks arrive most vividly when research follows emotions that surface during human–AI contact. Ricardo Santos, Amélia Brandão, Bruno Veloso and Paolo Popoli, in “The Use of AI in Government and its Risks: Lessons from the Private Sector,” mix sentiment analysis with qualitative reading of Reddit discussions and then reason carefully about public-sector transfer. Positive reactions appear frequently; negative reactions also appear in nontrivial proportions, including anger, frustration, fear and distrust that spill over onto organizational brands. That pattern matters for agencies that rely on citizen cooperation, since brand damage within public services erodes compliance, engagement and perceived fairness. The authors argue for staged deployment, frequent listening through sentiment tools, and proactive explanation that treats users as partners rather than passive recipients of automation. A public agency that introduces chatbots or decision aides without transparent communication, appeal rights and staff backstops invites wider reputational costs that outlast any single project.Urban governance presents a different canvas, where data, sensors and service platforms meet everyday life. Gerardo Bosco, Vincenzo Riccardi, Alessia Sciarrone, Raffaele D’Amore and Anna Visvizi, through “AI-Driven Innovation in Smart City Governance: Achieving Human-Centric and Sustainable Outcomes,” review international sources and synthesize managerial takeaways for planners and civic leaders by introducing a novel framework that identifies new measurement of the ethical impact of AI. Emphasis falls on design choices that protect inclusion, on governance models that mediate between municipal ambition and citizen rights, and on policy instruments that check enthusiasm against environmental, economic and social pillars. Their discussion tracks how AI-assisted services can amplify or diminish participation, depending on procurement rules, feedback loops, data standards and attention given to vulnerable groups. The framework is based on transparent coordination among actors, with metrics that capture social value, not only throughput or savings. These results open a debate around open government and co-production, while inflecting them with current algorithmic realities.Infrastructure projects highlight AI’s visibility for citizens who watch budgets, timelines and integrity signals during procurement and delivery. “Exploring the Impact of Artificial Intelligence on the Transparency and Rationality of Peruvian Public Works: Perceptions, Expectations, Challenges and Opportunities” assesses civil-works oversight and administrative discretion, while asking through structural equation modeling whether AI tools aid transparent practice and reasonable decision paths. Guillermo Miranda-Hospinal, José Miguel Villodre, David Valle-Cruz and Jorge L. Yrivarren-Lazo locate public-works oversight within broader debates about explainability, discretion and administrative rationality. Evidence and argument situate Peru’s institutional context while speaking broadly about design rules that favor auditability and citizen trust. That conversation aligns with international scholarship on algorithmic transparency and AI capability within administrations, and it brings construction oversight, a sensitive and visible sector, squarely into current AI governance dialogue.Across the contributions of the Special Issue several themes recur. Literacy emerges as a public competence that shapes expectations and behavior toward automated services; news narratives about generative systems frame those expectations daily, which strengthens the case for collaboration between media organizations and regulators. Responsible deployment appears as a managerial craft; teams management require time, staff learning and feedback loops that capture impacts across equity, privacy, accuracy and experience. Emotions can help managers see early evidence of confusion or anger that often precedes policy complaints or disengagement. Urban initiatives show how technical ambition must sit inside inclusive planning, otherwise smart programs risk leaving behind residents who already face access barriers. Education research confirms that perceived value drives use and gendered patterns, while not uniform, require targeted support that expands capability without stigmatization. Procurement and infrastructure add a call for transparency that regular citizens can recognize, including logs, explanations and open data that let journalists, auditors, and communities track decisions.From a methodological standpoint, the wide adoption of mixed-method designs reveals that purely technical evaluations rarely capture phenomena in-depth. For instance, the Reddit-based analysis treats user chatter as early warning, while qualitative reading grounds sentiment categories in context rather than abstraction. City-governance synthesis reads across disciplines, provides pathway recommendations for administrators and resists technological hype by insisting on social outcomes as evaluative baselines. Survey-based modelling in the paper on education advances a clean theory test with transparent diagnostics, which practitioners can translate into program design without heavy jargon. The article on media literacy leverages a real news corpus that ordinary people already consume, which strengthens external validity for proposed literacy components. Together, those methodological choices build a chain from everyday narratives through managerial action up toward formal policy.The varied methodological nature of the works according to multidimensional approaches is also reflected in the content. To identify a common thread among the papers of the Special Issue from an ecosystemic perspective, it can be noticed that the issue of generative AI cannot be analyzed only from a technical and legal-institutional point of view, but it is necessary to identify the drivers in terms of skills and strategic management useful for guiding the regulatory decisions and policies.The first article (Jitendra K. Pandey) immediately clarifies the need for AI implementation to be supported by a strategic orientation based on continuous learning, therefore on training human resources and adapting skills before demanding a fruitful use of these technological systems in public organizations.In the second article (Yucong Lao and Yukun You), the role of government as coordinators in the relationship with stakeholders confirms the need for an ecosystems-integrated approach based on the collaboration between governments, local authorities, civil society organizations and the private sector (Monda et al., 2023), for example, on co-governance.The third article (Testa, Mario Testa, Maddalena Della Volpe, Antonio D’Amato and Adriana Apuzzo) reveals the necessity of more inclusive educational and learning models, personalized to different needs, to address gendered digital divide.By focusing on Digital Humanism, the fourth article (Ricardo Santos, Amélia Brandão, Bruno Veloso, and Paolo Popoli) explores the different facets of human–computer interactions by revealing the emotional risks related to AI, in terms of anxiety and fear due to the higher rigidity of automated systems when compared to humans.The suitability of a multi-layered perspective to the issue is confirmed in the fifth article (Gerardo Bosco, Vincenzo Riccardi, Alessia Sciarrone, Raffaele D’Amore and Anna Visvizi) that seeks to introduce a common language and unified criteria for an ethical use of generative AI, combining human and societal dimensions (well-being, literacy) with technical features (data management).By analyzing the issue of control and monitoring, Miranda-Hospinal, José Miguel Villodre, David Valle-Cruz and Jorge L. Yrivarren-Lazo reveal that generative AI can improve traceability and transparency only through the integration of civic access.Hence, drawing a synthetic vision, critically incorporated into the framework proposed in Figure 1, generative AI use should start from orientation (grounded on social values, ethics and continuous learning) at a strategic level (the top of the pyramid) and then be translated into related human management and knowledge management strategies for digital literacy and the acquisition of a data-driven approach (at a functional level, the center of the pyramid) until it is then applied to real operations and processes in technical terms (at an operational, the basis of the pyramid).At a strategic level, the studies reveal the need of a participatory method to involve the public in AI policy and that governance frameworks societal through and collaboration for the optimization of government human–computer interactions should be et al., through skills (Visvizi, to data-driven orientation et al., which is reflected in a related data analysis design to support policymaking et al., the real implementation of tools, systems and algorithms the monitoring of impact during processes to and security of data while transparency and through citizens’ framework from the of the results and implications of the papers included in the Special Issue shows rather than technical government should into deeper societal and ethical in the adoption of generative AI by governance frameworks based on et al., Moreover, the enrichment of digital literacy is but only with the monitoring of and emotional reactions of users and citizens to AI follows while against should that include evaluation of time and to when Procurement language should data and while room for with service and risk teams to sentiment across channels, not as a but as an early for service design and where in leaders should explanations in language and add human service until confidence programs should algorithmic tools with human and that and external City should AI projects with rules, and participatory that residents into rather than as a Education and should staff development that capability across teaching including that use and research from gaps that authors media studies test whether literacy frameworks across broadcasters and that work would help global without a single that strategy research should process metrics and impacts in while governance that learning rather than risk monitoring would benefit from across strategies and careful attention to appear or public agencies should not service design around early while citizens who face language or barriers. scholarship can measurement beyond and while that a single host of that digital Education studies should results in different and with that test and ethics designs would help from structural oversight invites case studies that track AI-assisted monitoring from procurement through drawing on that include community and this issue as a contribution for and administrators who A first thread concerns with of these papers claims each show how careful design and listening while A second thread concerns literacy in audiences interpret new tools, students within and citizens which research must understanding as a social process that policy can A third thread concerns and damage within public services teams must watch emotions with the same for cost or throughput A fourth concerns and transparency that auditors can use appeal not A fifth concerns staged with evaluation agencies from and create room for community to around tools that real without of these and implications into a framework, like the in Figure or by Gerardo Bosco, Vincenzo Riccardi, Alessia Sciarrone, Raffaele D’Amore and Anna Visvizi in can the of a unified with for the impact of AI et al., the of negative emotions with AI, such as social distrust or can help government the that hinder a of these and the drivers for the of technology the different of to address gender digital public organizations can more detect the skills and for the development of agencies face algorithmic while communities for and value from digital This Special Issue from that A single cannot policy the set for with from contexts, and implications from each