Dissertation Proposal Table of Chapter 1: Introduction. 1 Statement of the Problem.. 4 Purpose of the Study. 5 Introduction to Framework. 6 Introduction to Research Methodology and Design. 8 Research Question. 8 Hypotheses. 9 Significance of the Study. 10 Definitions of Key Terms. 11 Summary. 12 Chapter 2 Literature Review.. 13 Documentation. 14 Read More
Visit the URL and place your order with us. Fill basic details of your research paper, set the deadlines and submit the form.
Chat with our experts to get the best quote. Make the payment via online banking, debit/credit cards or through paypal. Recieve an order confirmation number.
Sit back and relax. Your well written, properly referenced research paper will be mailed to your inbox, before deadline. Download the paper. Revise and Submit.
Dissertation Proposal
Table of
Introduction to Research Methodology and Design. 8
Chapter 2 Literature Review.. 13
Artificial intelligence (AI) has become a driving force behind digital transformation in government organizations, reshaping how public services are delivered, managed, and regulated. As local and state agencies increase their use of cloud-based AI tools for automation, decision support, and operational efficiency, new governance challenges have emerged. These challenges extend beyond technical risk considerations to encompass expectations regarding transparency, accountability, and regulatory compliance in environments where decisions increasingly depend on complex algorithmic systems (Folorunso et al., 2024; Lothery, 2024; Robles, 2023). While the adoption of AI continues to accelerate, the development of integrated governance mechanisms has lagged, creating uncertainty about how government entities will maintain oversight, public trust, and statutory alignment in AI-enabled operations (Robles, 2023). These dynamic underscores the need for research examining how government professionals understand, interpret, and evaluate the frameworks intended to guide responsible and accountable AI adoption.
Government organizations operate within regulatory ecosystems that require rigorous accountability processes, standardized documentation, and strict adherence to statutory requirements. Unlike private-sector institutions, public agencies must manage AI deployment in ways that meet expectations for transparency, ethical conduct, and public trust, while also complying with federal, state, and local mandates. Cloud service providers now play a central role in delivering AI capabilities, embedding these tools into mission-critical systems. This growing dependence underscores the need for governance models that clearly define responsibilities, compliance expectations, and mechanisms to ensure operational resilience. Yet the rising complexity of AI and cloud infrastructures has made it difficult for agencies to adopt governance models that are both comprehensive and practical (Galij et al., 2024; Jansen et al., 2023).
To address these challenges, standards bodies have introduced conceptual frameworks to guide responsible AI risk management. Among the most prominent is the National Institute of Standards and Technology (NIST) Artificial Intelligence Risk Management Framework (AI RMF 1.0), which outlines principles for promoting trustworthy AI systems and improving organizational governance capabilities (NIST, 2023). The framework emphasizes transparency, accountability structures, and compliance with broader regulatory expectations. However, recent studies indicate that government entities continue to struggle with applying these principles in operational settings. Researchers highlight inconsistent interpretations of governance guidance, fragmented oversight structures, and difficulty translating conceptual principles into practice-oriented processes (Folorunso et al., 2024; Lothery, 2024; Robles, 2023). These challenges are especially evident in local government agencies, where formal AI governance structures may be underdeveloped and reliance on external cloud providers is high.
Although scholarship examined AI ethics, transparency, and risk management, empirical research has not evaluated how government practitioners perceive the effectiveness of AI governance frameworks. Much of the existing literature focuses on theoretical models, national-level oversight, or policy recommendations rather than on practitioner-level experiences in local government (Agarwal & Nene, 2025; Schmitz et al., 2025). The emergence of “agentic AI” systems capable of autonomous decision-making adds additional layers of complexity, raising concerns about oversight demands, auditability, interpretability, and organizational accountability (Schmitz et al., 2025). These factors reinforce the need for empirical inquiry into how government professionals assess governance frameworks and whether these frameworks sufficiently support the responsible adoption of AI technologies in practice.
The Technology Acceptance Model (TAM) provides a sound theoretical foundation for investigating these dynamics. Initially introduced by Davis (1986, 1989), TAM proposes that perceived usefulness and perceived ease of use shape user acceptance and adoption of new technologies. Later extensions incorporated constructs related to trust, transparency, and organizational context, reflecting the increasing need to align technology adoption with governance expectations (Holden & Karsh, 2010; Venkatesh & Bala, 2008; Venkatesh & Davis, 2000). TAM has been widely applied in public-sector research, particularly in studies examining the adoption of information systems, cybersecurity processes, and regulatory frameworks (Wirtz et al., 2023). Its structure makes it well suited to understanding how government professionals evaluate key characteristics of AI governance frameworks, such as transparency, accountability, and clarity of compliance.
Within the context of AI governance, TAM’s core constructs map directly onto the characteristics emphasized in the NIST AI RMF 1.0. Perceived usefulness aligns with professionals’ beliefs about the extent to which the framework enhances oversight, improves risk management, and strengthens accountability. Perceived ease of use corresponds to perceptions about the clarity and interpretability of compliance requirements. Transparency and accountability, which play central roles in responsible AI governance, also influence trust and confidence factors identified in TAM literature as shaping perceived usefulness and ease of use (Holden & Karsh, 2010; Venkatesh & Bala, 2008). These relationships provide the conceptual foundation for examining how perceptions of transparency, accountability, and compliance clarity influence broader evaluations of AI RMF effectiveness in local government settings. TAM also informs the proposed research design by identifying the perceptual variables that shape adoption behavior. By examining how government professionals interpret specific characteristics of the NIST AI RMF 1.0, the study will link theoretical propositions to real-world practitioner experiences. This approach supports both theoretical advancement and practical insights by extending TAM into the domain of AI governance.
As public-sector agencies face increasing technical, regulatory, and ethical pressures surrounding AI adoption, understanding how practitioners evaluate governance frameworks becomes essential. Such insights can strengthen oversight processes, enhance compliance strategies, and ultimately improve the trustworthiness and accountability of AI-enabled government operations. The introduction presented here establishes the study’s broader context, theoretical grounding, and scholarly relevance. The following section transitions to the specific problem addressed by this research, advancing the progression toward an empirical examination of how AI governance frameworks function within local government organizations.
The problem to be addressed in this study is the lack of a comprehensive, practice-based governance framework that public-sector organizations can apply to ensure compliance when adopting artificial intelligence (AI) from cloud providers (Folorunso et al., 2024; Lothery, 2024; Robles, 2023). Without an accepted recognizable framework, government agencies face increasing risks tied to fragmented oversight structures, inconsistent application of compliance standards, and reliance on external cloud providers for critical AI services. These shortcomings have undermined organizational accountability, created vulnerabilities in regulatory compliance, and weakened resilience against geopolitical, technical, and operational disruptions (Eisenberg et al., 2025).
Researchers have highlighted that the adoption of “agentic AI” systems by public organizations has intensified governance challenges, particularly in oversight, continuous monitoring, and interdepartmental coordination (Schmitz et al., 2025). Similarly, Agarwal and Nene (2025) argued that governance remains fragmented across regulatory, standards, and domains, leaving government organizations without coherent guidance for practice. This creates not only compliance gaps but also accountability challenges for policymakers and IT leaders charged with safeguarding public trust in AI-enabled services. Without accountability standards, third-party audits cannot consistently ensure compliance.
The existing literature provides limited discussion of the mechanisms of accountability for government stakeholders and of how the public responds to governance challenges. While conceptual frameworks exist, they remain fragmented and lack the empirical grounding necessary for practical applicability in government contexts (Folorunso et al., 2024). Unless scholars and practitioners develop a governance approach that incorporates sovereignty, adaptive compliance, and operational resilience, government organizations will continue to face barriers in safely and effectively leveraging AI cloud services.
The purpose of this quantitative correlational study is to examine the relationship between perceived transparency, perceived accountability, and perceived compliance clarity (independent variables) and the perceived effectiveness of the NIST AI RMF 1.0 (dependent variable) within local Sacramento County government agencies. Perceived transparency refers to the extent to which government professionals believe the framework clearly communicates risks, processes, and decision-making criteria for AI governance (Folorunso et al., 2024). Perceived accountability reflects participants’ beliefs that the NIST AI RMF 1.0 establishes clear roles, responsibilities, and oversight mechanisms for AI adoption (Agarwal & Nene, 2025; Schmitz et al., 2025). Perceived compliance clarity is defined as the extent to which practitioners perceive the framework’s compliance expectations as understandable and actionable for ensuring regulatory adherence (Lothery, 2024; Robles, 2023).
A sample size of 182 structured electronic surveys will be sent to Sacramento County government professionals involved in AI or cloud governance. Using Likert-scale items ranging from 1 (Strongly Disagree) to 5 (Strongly Agree), the survey measures perception of transparency, accountability, compliance clarity, and the overall effectiveness of the framework. Descriptive statistics, Pearson correlations, and multiple linear regression will be used to examine whether the three independent variables significantly predict perceived effectiveness of the NIST AI RMF 1.0. A target sample, as illustrated in Figure 1, of approximately 182 participants was identified through a priori power analysis using G*Power 3.1 for multiple linear regression with three predictors. Assuming a small-to-moderate effect size (f² = 0.08), an alpha level of .05, and power of .90, a minimum of 182 participants is required (Cohen, 1988; Faul et al., 2009).
This study is guided by the Technology Acceptance Model (TAM), initially proposed by Davis (1986, 1989) and later expanded by Venkatesh and Davis (2000). TAM is one of the most widely applied frameworks for predicting and explaining user adoption of new technologies. It suggests that two key cognitive beliefs, perceived usefulness (PU) and perceived ease of use (PEOU), influence a user’s attitude toward using a system, which subsequently affects their behavioral intention to use, and ultimately, their actual use of the system. Perceived usefulness reflects the extent to which an individual believes that using a system will enhance their job performance, while perceived ease of use represents the degree to which a person believes that using the system will be free of effort (Davis, 1989). These constructs collectively explain how users form acceptance decisions toward emerging technologies (Venkatesh & Davis, 2000).
TAM has been extensively validated and adapted in studies examining the adoption of information systems, cybersecurity frameworks, and governance mechanisms in public-sector contexts (Holden & Karsh, 2010; Wirtz et al., 2023). In this study, TAM provides the conceptual foundation for understanding how government professionals’ perceptions of transparency, accountability, and clarity of compliance influence their perceived effectiveness of the NIST Artificial Intelligence Risk Management Framework (NIST AI RMF 1.0). Within this context, perceived usefulness refers to the extent to which agency leaders and compliance officers believe that the NIST AI RMF 1.0 improves oversight, risk management, and accountability in AI adoption. Perceived ease of use is consistent with the construct of compliance clarity, reflecting how readily professionals can interpret and apply the framework’s guidelines. Additionally, transparency and accountability enhance users’ trust and confidence in the system, indirectly shaping both perceived usefulness and intention to sustain framework adoption (Venkatesh & Bala, 2008).
The central propositions derived from TAM and adapted for this study are as follows: Perceived transparency is expected to positively influence the perceived usefulness of the NIST AI RMF 1.0, while perceived accountability is anticipated to strengthen perceived usefulness by enhancing trust and organizational confidence. Additionally, perceived compliance clarity is posited to positively affect perceived ease of use, thereby improving perceptions of framework effectiveness. Ultimately, both perceived usefulness and perceived ease of use are theorized to collectively predict the perceived effectiveness and sustained adoption of the NIST AI RMF 1.0.
This study employs a quantitative correlational design to examine relationships between perceived transparency, accountability, and compliance clarity, and their collective influence on the perceived effectiveness of the NIST AI Risk Management Framework (AI RMF 1.0) among local government professionals. The quantitative approach enables objective measurement and statistical analysis of these constructs, aligning with the study’s purpose to determine predictive relationships rather than causal effects (Creswell & Creswell, 2018; Field, 2018).
Data will be collected using a structured Likert-scale survey, distributed electronically to local government participants. Survey items, adapted from established Technology Acceptance Model (TAM) instruments (Davis, 1989; Venkatesh & Davis, 2000) and governance studies (Holden & Karsh, 2010; Wirtz et al., 2023), will measure perceptions of transparency, accountability, compliance clarity, and framework effectiveness on a five-point scale ranging from strongly disagree (1) to strongly agree (5).
Data analysis will include descriptive statistics to summarize participant characteristics, followed by Pearson correlation and multiple regression analyses to assess the strength and direction of relationships among variables. Statistical significance will be tested at α = 0.05. This design directly supports the problem, purpose, and research questions by quantifying how perceptions of framework characteristics influence the perceived effectiveness and adoption of the NIST AI RMF 1.0 within government contexts.
To what extent does a statistically significant correlation exist between perceived effectiveness among participants and the effectiveness of NIST AI RMF 1.0?
To what extent does a statistically significant correlation exist between perceived accountability among participants and the effectiveness of NIST AI RMF 1.0?
To what extent does a statistically significant correlation exist between perceived compliance clarity among participants and the effectiveness of NIST AI RMF 1.0?
There is no statistically significant relationship between local government professionals’ perceptions of transparency and their perceptions of the effectiveness of the NIST AI RMF 1.0.
There is a statistically significant relationship between local government professionals’ perceptions of transparency and their perceptions of the effectiveness of the NIST AI RMF 1.0.
There is no statistically significant relationship between local government professionals’ perceptions of accountability and their perceptions of the effectiveness of the NIST AI RMF 1.0.
There is a statistically significant relationship between local government professionals’ perceptions of accountability and their perceptions of the effectiveness of the NIST AI RMF 1.0.
There is no statistically significant relationship between local government professionals’ perceptions of compliance clarity and their perceptions of the effectiveness of the NIST AI RMF 1.0.
There is a statistically significant relationship between local government professionals’ perceptions of compliance clarity and their perceptions of the effectiveness of the NIST AI RMF 1.0.
This study is significant for advancing understanding of how perceived transparency, accountability, and clarity of compliance influence the perceived effectiveness of the NIST AI Risk Management Framework (AI RMF 1.0) within local government organizations. As public-sector agencies increasingly adopt artificial intelligence through cloud platforms, ensuring effective governance and accountability mechanisms becomes crucial for the ethical and compliant implementation of this technology. The study contributes to the literature by extending the Technology Acceptance Model (TAM) into the domain of AI governance, offering an empirically validated framework that links user perceptions to governance outcomes (Davis, 1989; Venkatesh & Davis, 2000).
From a theoretical perspective, the findings will enrich the TAM framework by integrating governance-related constructs, such as transparency, accountability, and compliance clarity, within the context of federal and local government technology adoption. For practitioners, the study offers actionable insights for policymakers, Chief Information Officers (CIOs), and compliance leaders to enhance adoption strategies and improve the perceived usefulness and usability of the AI RMF 1.0. Addressing the study problem may lead to improved implementation fidelity, greater stakeholder trust, and more substantial alignment between NIST’s AI governance principles and real-world agency practices. Ultimately, achieving the study purpose and answering the research questions will contribute to more consistent, transparent, and accountable AI governance in the public sector, strengthening both ethical outcomes and operational resilience.
The obligation of organizations or individuals to justify decisions and actions regarding AI governance, ensuring transparency, ethical use, and responsibility across operational processes (Floridi & Cowls, 2021).
The ability of computer systems to perform tasks that normally require human intelligence, including learning, reasoning, and self-correction (Russell & Norvig, 2021).
The degree to which governance frameworks provide clear, interpretable, and actionable requirements that facilitate adherence to policies and regulatory standards (Wirtz et al., 2023).
A structured set of policies, processes, and standards designed to guide the ethical and compliant management of AI systems within an organization (Jansen et al., 2023).
The National Institute of Standards and Technology Artificial Intelligence Risk Management Framework is a voluntary framework providing principles and guidelines to manage risks associated with AI development and deployment (NIST, 2023).
A government entity responsible for delivering services, enforcing regulations, and managing public resources, often constrained by statutory and compliance requirements (Younus et al., 2025).
This chapter outlined the conceptual foundations, theoretical framework, and research design guiding this study. It introduced the application of the Technology Acceptance Model (TAM) to examine how perceived transparency, accountability, and clarity of compliance influence perceptions of the NIST AI RMF 1.0’s effectiveness in local government contexts. The discussion included an overview of the quantitative correlational design, data collection process, and analytical methods, followed by the study’s significance and key operational definitions. The next chapter will present a comprehensive review of the relevant literature, identify gaps in AI governance research, and situate this study within the broader context of public-sector technology adoption.
The problem to be addressed in this study is the lack of a comprehensive, practice-based governance framework that public-sector organizations can apply to ensure compliance when adopting artificial intelligence (AI) from cloud providers (Folorunso et al., 2024; Lothery, 2024; Robles, 2023). The purpose of this quantitative correlational study is to examine the relationship between perceived transparency, perceived accountability, and perceived compliance clarity (independent variables) and the perceived effectiveness of the National Institute of Standards and Technology Artificial Intelligence Risk Management Framework (NIST AI RMF 1.0) (dependent variable) within local Sacramento County government agencies.
Artificial intelligence has rapidly evolved from an emerging technology into a strategic capability supporting decision-making, service delivery, and operational efficiency across government organizations. Public agencies increasingly rely on AI-enabled cloud technologies to improve administrative processes, strengthen cybersecurity, analyze large volumes of data, and support evidence-based policymaking. While these technologies provide significant operational benefits, they also introduce governance challenges related to transparency, accountability, ethical oversight, regulatory compliance, and organizational risk management. Recent studies suggest that existing governance approaches have not evolved at the same pace as AI technologies, resulting in fragmented governance practices and inconsistent implementation across public-sector organizations (Almeida et al., 2022; Criado, 2025; Folorunso et al., 2024).
The rapid adoption of cloud-based AI has further intensified governance challenges by increasing organizational dependence on commercial cloud providers while simultaneously expanding concerns regarding digital sovereignty, data governance, regulatory compliance, and public trust. Researchers have argued that governments frequently adopt AI technologies without comprehensive governance mechanisms capable of ensuring transparency, accountability, and continuous oversight throughout the AI lifecycle (Jansen et al., 2023; Liu, 2025; Luitse, 2024). Although governance frameworks such as the NIST AI RMF 1.0 provide principles for trustworthy AI, empirical evidence regarding how government practitioners perceive these governance principles remains limited, particularly within local government organizations responsible for implementing AI-enabled cloud services (Robles, 2023; Schmitt, 2024).
Recent scholarship has increasingly emphasized that responsible AI governance extends beyond technical implementation to encompass organizational leadership, ethical decision-making, stakeholder engagement, and continuous governance throughout the AI lifecycle. Researchers consistently identify transparency, accountability, explainability, fairness, and compliance as foundational characteristics of trustworthy AI; however, they also acknowledge that organizations continue to experience challenges translating these governance principles into operational practice (Almeida et al., 2022; Liu, 2025; Vatamanu et al., 2025). Consequently, understanding how government professionals perceive these governance characteristics has become increasingly important for evaluating the effectiveness of AI governance frameworks and supporting responsible AI adoption within the public sector.
The literature reviewed in this chapter provides a comprehensive synthesis of research related to AI governance, ethical AI implementation, transparency, accountability, compliance clarity, and organizational readiness for AI adoption within public-sector organizations. The review begins by describing the documentation process used to identify relevant peer-reviewed literature and presents the Technology Acceptance Model (TAM) as the theoretical framework guiding the study. The chapter then synthesizes the literature through thematic discussions of AI governance and ethics in public-sector organizations, transparency and explainability in AI governance, accountability mechanisms, compliance clarity and the operationalization of governance frameworks, and organizational readiness for AI adoption. Finally, the chapter identifies the gaps in the existing literature that provide the foundation and justification for the present study.
The databases and search engines used for this literature review included Google Scholar, PubMed, and arXiv. These platforms were selected to ensure a diverse coverage of peer-reviewed scholarship, interdisciplinary research, and studies related to artificial intelligence (AI) governance in public-sector contexts. The advanced search string used in arXiv included: all:AI governance AND all:public sector AND all:ethics OR all:algorithmic accountability AND all:transparency AND all:fairness. The keywords used focused on capturing research addressing ethical oversight, accountability mechanisms, and transparency in government AI applications. A filter was added for items with a data range from 2020 to 2025.
Google Scholar was used to identify peer-reviewed journal articles, conference proceedings, doctoral dissertations, and policy analyses addressing AI governance frameworks, regulatory compliance, and public-sector adoption. The following Boolean search strategy was employed to capture relevant literature: ((“artificial intelligence”[Title/Abstract] OR “AI”[Title/Abstract] OR “machine learning”[Title/Abstract]) AND (“governance”[Title/Abstract] OR “governance framework”[Title/Abstract] OR “policy”[Title/Abstract]) AND (“public sector”[Title/Abstract] AND (“ethics”[Title/Abstract] OR “ethical”[Title/Abstract] OR “accountability”[Title/Abstract] OR “transparency”[Title/Abstract] OR “fairness”[Title/Abstract])). A filter was added for items with a data range from 2020 to 2025.
PubMed was searched using the same structured Title/Abstract Boolean query to ensure consistency across databases and to capture governance-related research in regulated public service environments. The following Boolean search strategy was used: ((“artificial intelligence”[Title/Abstract] OR “AI”[Title/Abstract] OR “machine learning”[Title/Abstract]) AND (“governance”[Title/Abstract] OR “governance framework”[Title/Abstract] OR “policy”[Title/Abstract]) AND (“public sector”[Title/Abstract] AND (“ethics”[Title/Abstract] OR “ethical”[Title/Abstract] OR “accountability”[Title/Abstract] OR “transparency”[Title/Abstract] OR “fairness”[Title/Abstract])). A filter was added for items with a data range from 2020 to 2025.
The conceptual framework guiding this section of the literature review is the Technology Acceptance Model (TAM), which explains how users form perceptions about and accept new technologies or systems (Davis, 1989; Venkatesh & Davis, 2000). TAM proposes that individuals’ adoption of a system is primarily influenced by two interrelated perceptions: perceived usefulness and perceived ease of use. These perceptions shape attitudes toward adoption and the likelihood that individuals will integrate a tool or framework into practice. In the context of AI governance in public-sector organizations, TAM provides a useful lens for examining how practitioners evaluate governance frameworks intended to guide responsible AI deployment. Specifically, constructs such as transparency, accountability, and compliance clarity can be interpreted as factors that influence perceived usefulness and perceived ease of use, thereby shaping how government professionals assess the operational effectiveness of governance frameworks such as the NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0).
Perceived compliance clarity is the extent to which the framework’s requirements are understandable and actionable (Schmitz et al., 2025). In TAM terms, this variable is a direct precursor to “perceived ease of use.” If the NIST AI RMF is seen as overly abstract or fragmented, the “perceived ease of use” will be low, leading to resistance or superficial implementation (Dotan et al., 2024). Therefore, frameworks that translate governance principles into clearly defined controls, processes, and accountability mechanisms are likely to be adopted.
Compliance clarity is particularly critical in the public sector, where strict statutory requirements bind agencies and often operate with limited resources (Jansen et al., 2023). A framework that provides a “lean control catalog” and clear mappings to existing cybersecurity standards is more likely to be perceived as easy to use, thereby improving its overall perceived effectiveness (Hussain et al., 2024). When governance requirements are clearly defined and operationalized, public-sector leaders are likely to integrate them into an existing risk management and compliance workflow.
In this study, perceived accountability may influence how public-sector practitioners evaluate the effectiveness of governance frameworks, such as the NIST Artificial Intelligence Risk Management Framework (AI RMF). The AI RMF emphasizes transparency, responsibility, and oversight as foundational principles for trustworthy AI implementation. As a result, practitioners who perceive stronger accountability mechanisms may view the framework as more effective in guiding ethical decision-making, managing risk, and supporting public trust in AI-enabled systems. Legal-policy alignment ensures that governance frameworks are adopted in a vacuum that integrates them with existing statutory and regulatory requirements. Ethical design principles provide a normative baseline for what “trustworthy” means in practice, moving beyond mere technical compliance. Technical auditability enables verification of governance claims, turning “perceived transparency” into verifiable data. Finally, multi-stakeholder engagement ensures that the governance process is inclusive, incorporating the diverse perspectives of developers, practitioners, and citizens, thereby reinforcing the “public value” created by the framework (Li, 2025; Palamarchuk, 2024; Pasupuleti, 2025).
Perceived compliance clarity referred to the extent to which the requirements of a governance framework are understandable, interpretable, and actionable for practitioners responsible for implementation (Schmitz et al., 2025). Within the Technology Acceptance Model (TAM), compliance clarity serves as a precursor to perceived ease of use, as practitioners were more likely to adopt frameworks whose requirements can be translated into concrete operational tasks. When governance guidance is abstract or fragmented, practitioners often struggle to determine how policies map to technical controls or operational processes, resulting in partial or inconsistent implementation (Dotan et al., 2024). This challenge is particularly pronounced in AI governance frameworks, where regulatory language frequently describes high-level ethical or accountability goals without specifying measurable implementation practices.
Compliance clarity is especially critical in the public sector because agencies must demonstrate adherence to statutory mandates while operating with constrained resources and complex oversight environments (Jansen et al., 2023). Explicit mappings to established cybersecurity standards, such as the NIST Cybersecurity Framework, are therefore more likely to be perceived as usable and effective by public-sector practitioners (Hussain et al., 2024). Research on AI-enabled compliance systems further suggests that automation technologies can help operationalize regulatory frameworks by translating textual policy requirements into measurable technical behaviors and monitoring mechanisms. For example, generative AI systems can analyze regulatory documents using natural language processing and map those requirements to system logs, security controls, and operational configurations, thereby reducing the reliance on manual interpretation of policy language (Yang et al., 2023; Sharma, 2024). Such capabilities demonstrate how emerging technologies may enhance perceived compliance clarity by bridging the gap between policy interpretation and technical implementation.
Recent studies suggested that generative AI could shift compliance management from a reactive, documentation-driven approach to a more proactive, continuously monitored governance process (Yang et al., 2023; Sharma, 2024). By leveraging techniques such as large language models, vectorized knowledge bases, and retrieval-augmented generation (RAG), organizations can automatically interpret regulatory clauses, generate risk assessments, and provide contextual mitigation recommendations when compliance gaps are detected. These systems could analyze complex regulatory texts, identify deviations from compliance standards, and generate actionable remediation guidance in real time (Yang et al., 2023; Kandpal et al., 2023). In regulated domains such as healthcare, for example, generative AI architectures have been proposed to dynamically map operational telemetry to compliance rules, including HIPAA safeguards and NIST cybersecurity practices, producing automated compliance scorecards and continuous monitoring insights.
The proposed conceptual framework for this study suggests that the effectiveness of the NIST AI Risk Management Framework (AI RMF) is supported by four foundational governance pillars identified in the public-sector AI governance literature: (1) legal-policy alignment, (2) ethical design principles, (3) technical auditability, and (4) multi-stakeholder engagement (Pasupuleti, 2025). Legal-policy alignment ensures that governance frameworks are embedded within existing statutory and regulatory structures rather than operating independently of established compliance requirements. Ethical design principles provide a normative baseline for responsible AI development, translating abstract concepts such as fairness and accountability into system design considerations. Technical auditability enables verification of governance claims by generating traceable evidence of compliance activities, transforming perceived transparency into measurable data. Finally, multi-stakeholder engagement strengthens governance legitimacy by incorporating the perspectives of policymakers, developers, practitioners, and citizens, thereby reinforcing the public value created by AI governance initiatives (Li, 2025; Palamarchuk, 2024; Pasupuleti, 2025).
Pasupuleti (2025) examined governance frameworks designed to support the responsible deployment of artificial intelligence within public-sector services. The study emphasized that as governments increasingly integrate AI into administrative decision-making and digital service delivery, structured governance mechanisms are necessary to ensure AI systems operate transparently, accountably, and in compliance with the law. Through comparative analysis of public-sector AI initiatives, the research identified four core governance components that supported effective AI oversight: legal-policy alignment, ethical design principles, technical auditability, and multi-stakeholder engagement.
Legal-policy alignment ensures that AI systems operate within established regulatory and statutory requirements, while ethical design principles guide the embedding of fairness, transparency, and accountability in AI development processes. Technical auditability enables organizations to verify and monitor AI system behavior through documentation, traceability, and evaluation mechanisms, thereby validating governance claims with measurable evidence. In addition, multi-stakeholder engagement strengthens governance legitimacy by incorporating the perspectives of policymakers, technologists, and citizens in AI oversight processes. These governance pillars provided a conceptual foundation for evaluating AI governance frameworks, such as the NIST AI Risk Management Framework, and support the present study’s focus on perceived transparency, accountability, and compliance clarity as factors influencing the perceived effectiveness of AI governance in public-sector organizations.
Kim et al. (2024) the institutional conditions required for the adoption of artificial intelligence systems in the public sector, emphasizing accountability as a central governance requirement for responsible AI deployment. The authors argued that although AI technologies offer significant potential to improve government decision-making and service delivery, public-sector adoption must be evaluated through the lens of institutional accountability rather than technological capability alone. Drawing from public administration theory, the study used Romzek and Ingraham’s accountability framework to describe how hierarchical, legal, professional, and political accountability structures shape the governance of AI systems within government organizations. These accountability relationships determined how responsibility for decisions influenced by automated systems is assigned and how oversight mechanisms were implemented to ensure transparency and legitimacy in administrative decision-making (Kim et al., 2024).
From this author’s perspective, these governance mechanisms are closely related to the concept of perceived accountability, which refers to practitioners’ perceptions that clear responsibility, oversight, and reporting mechanisms exist for AI-supported decisions. When public-sector organizations establish well-defined accountability structures, practitioners were more likely to perceive that AI systems operate within appropriate governance boundaries and remain subject to human oversight. The authors further argue that AI systems could help reduce institutional uncertainty in decision-making by providing data-driven insights that improve administrative knowledge and policy implementation. However, the adoption of AI systems must be accompanied by governance structures that monitor algorithmic behavior and ensure that human actors remain accountable for policy outcomes (Kim et al., 2024).
These findings support the broader argument that accountability mechanisms play a critical role in shaping how government organizations evaluate the legitimacy and effectiveness of AI governance frameworks (Romzek & Ingraham, 2000; Schmitt, 2024). In this study, perceived accountability may influence how public-sector practitioners evaluate the effectiveness of governance frameworks, such as the NIST Artificial Intelligence Risk Management Framework (AI RMF). The AI RMF emphasizes transparency, responsibility, and oversight as foundational principles for trustworthy AI implementation (National Institute of Standards and Technology [NIST], 2023).
Romzek and Ingraham (2000) provided a foundational framework for understanding accountability within public administration, emphasizing that government organizations operate within multiple overlapping accountability relationships. The authors argue that accountability in the public sector is not a singular concept but rather a multidimensional system shaped by institutional structures, organizational hierarchies, legal mandates, professional norms, and political oversight. Their study analyzed the Ron Brown plane crash investigation to demonstrate how competing accountability pressures can influence decision-making and organizational behavior within government agencies.
Through this case analysis, the authors illustrated how public administrators often face cross-pressures from multiple accountability sources that may conflict, creating challenges for effective governance and oversight. Romzek and Ingraham (2000) identified four primary forms of accountability that guide public-sector decision-making: hierarchical, legal, professional, and political accountability. Hierarchical accountability operates within formal organizational structures, where administrators are responsible to supervisors and agency leadership through established chains of command. Legal accountability refers to oversight through statutory requirements, regulatory frameworks, and judicial review mechanisms that ensure government agencies operate within the law. Professional accountability is derived from the expertise and ethical standards of professional communities, specialized knowledge and norms guide decision-making in complex policy environments. Finally, political accountability arises from democratic governance structures in which elected officials and citizens hold government agencies responsible for policy outcomes and administrative actions.
The authors emphasized that effective public administration requires balancing these different accountability relationships, as reliance on a single oversight mechanism may lead to governance failures or diminished organizational effectiveness. In highly complex administrative environments, such as those involving advanced technologies or policy innovation, public officials must navigate accountability pressures while maintaining transparency and accountability in decision-making. The framework proposed by Romzek and Ingraham has since become influential in public administration research, highlighting the institutional dynamics that shape how government organizations implement policies and manage risk.
In the context of emerging technologies such as artificial intelligence, the multidimensional nature of accountability identified by Romzek and Ingraham provided an important theoretical foundation for understanding governance challenges in public-sector technology adoption. AI systems introduce new layers of complexity into administrative decision-making, including questions about algorithmic transparency, responsibility for automated decisions, and oversight of technical systems. Applying Romzek and Ingraham’s accountability framework to AI governance suggests that public-sector organizations must establish governance structures that clarify responsibility across hierarchical leadership, legal compliance requirements, professional expertise, and political oversight. These accountability mechanisms may shape practitioners’ perceptions of the legitimacy and effectiveness of governance frameworks designed to guide responsible AI implementation in government organizations.
Li and Sun (2025) examined the extent to which algorithmic accessibility alone is sufficient to foster trust in automated decision-making within governance contexts. The authors argue that while accessibility, defined as users’ ability to understand and interact with algorithmic systems, is an important component of AI governance, it is insufficient on its own to ensure trust. Instead, accountability emerges as a critical corresponding factor that shapes how individuals evaluate and accept automated decisions in public-sector environments.
The study conceptualized accessibility as a mechanism that enhances transparency by enabling users to interpret algorithmic processes and outputs. However, Li and Sun (2025) emphasized that transparency without accountability may fail to produce meaningful trust outcomes. Accessibility can improve user comprehension, but without clear lines of responsibility and mechanisms for redress, users may still perceive AI systems as unreliable or unjust. This distinction highlights a key limitation in many existing AI governance approaches, which often prioritize explainability while underemphasizing accountability structures.
Methodologically, the authors employed observational research design to assess relationships among accessibility, accountability, and trust in automated decision-making. Their findings indicate that accountability plays a facilitating and expanding role, strengthening the positive effects of accessibility on trust. When users perceive that decision-makers can be held responsible for algorithmic outcomes, their confidence in automated systems increases significantly. On the contrary, in the absence of perceived accountability, improvements in accessibility yield only marginal gains in trust (Li & Sun, 2025).
A significant contribution to the study is its introduction of accountability as a perceptual construct rather than solely an institutional or regulatory requirement. This perspective aligns closely with emerging literature in public-sector AI governance, where perceived accountability influences user acceptance, trust, and system legitimacy. From a theoretical standpoint, the findings can be situated within the Technology Acceptance Model (TAM), where accountability enhances perceived usefulness and trust, ultimately shaping adoption behavior. This reinforces the importance of integrating governance mechanisms that are not only technically robust but also visible and understandable to stakeholders.
The implications of this study were particularly relevant for public-sector organizations implementing AI systems through cloud-based platforms. Governance frameworks were required to extend beyond technical transparency to include enforceable accountability mechanisms such as audit trails, oversight structures, and clear assignment of responsibility. These findings support the principles outlined in the National Institute of Standards and Technology AI Risk Management Framework, which emphasizes accountability, transparency, and traceability as foundational elements of trustworthy AI. Without these mechanisms, public agencies risk deploying AI systems that fail to achieve stakeholder trust or meet compliance expectations.
From a critical perspective, the study’s strength lies in its empirical validation of the relationship between accessibility, accountability, and trust, providing measurable insights that extend beyond purely conceptual discussions. However, the research may be limited by its contextual scope, as perceptions of accountability can vary across established environments and cultural settings.
Adepoju and Chinonyerem (2025) examined the evolving role of algorithmic governance in the United States public sector, with a particular emphasis on designing transparent and accountable artificial intelligence (AI) systems for policy decision-making. The study addresses a critical gap in public-sector AI adoption: as reliance on algorithmic systems increases, tensions arise between efficiency gains and the preservation of democratic values such as transparency, fairness, and accountability. As AI becomes embedded across criminal justice, social welfare, and healthcare, the authors argue that governance frameworks must evolve to address the opacity and complexity of algorithmic decision-making.
A central contribution of the article is its identification of transparency as a foundational requirement for maintaining public trust and legitimacy in AI-driven governance. However, the authors highlighted that full technical transparency was often impractical due to the complexity of machine learning systems. Instead, the concept of “reasoned transparency,” which focused on providing understandable justifications for decisions rather than full system disclosure, was proposed as a more feasible approach. This aligned with the broader literature on explainable AI (XAI), which seeks to balance interpretability with system performance and security constraints.
Methodologically, the study employed a systematic literature review (SLR) guided by PRISMA standards, synthesizing 42 high-quality studies from an initial pool of 512 records. The analysis identified three dominant thematic clusters: (1) the importance of explainability and interpretability in fostering trust, (2) the need for robust institutional accountability mechanisms, and (3) the role of oversight structures such as auditing, ethical guidelines, and governance frameworks in ensuring responsible AI adoption. Adepoju and Chinonyerem (2025) noted that transparency and explainability account for the largest proportion of research focus (38.1%), followed by accountability (26.2%) and oversight mechanisms (21.4%), reinforcing the centrality of these constructs in AI governance.
The findings further emphasize that existing public-sector accountability mechanisms are insufficient to address the dynamic, autonomous nature of AI systems. Traditional approaches, such as periodic audits or ex-ante approvals, are described as fragmented and reactive. Instead, the authors advocated for continuous, lifecycle-based oversight and cross-departmental accountability structures that integrate auditing, risk assessment, and institutional review processes. This aligns with emerging governance approaches that emphasize ongoing monitoring and adaptive risk management rather than static compliance models.
Another significant contribution of the study is its exploration of sociotechnical challenges in AI governance. The authors argued that purely technical solutions, such as fairness metrics or explainability tools, are insufficient to address broader issues of equity and legitimacy. Instead, AI governance must be embedded within institutional, legal, and social contexts to ensure alignment with democratic principles. This perspective reflects the importance of integrating technical controls with organizational practices and policy frameworks, a key tenet of socio-technical systems theory.
From a policy perspective, the study identified several key governance enablers, including the adoption of national frameworks, the establishment of institutional oversight bodies, and the implementation of participatory transparency mechanisms. The authors reference initiatives such as the National Institute of Standards and Technology AI Risk Management Framework and the Blueprint for an AI Bill of Rights as foundational efforts to guide responsible AI adoption in the United States. However, they noted that implementation remains uneven across agencies, underscoring the need for standardized, enforceable governance practices.
Critically, the study’s strength lies in its comprehensive synthesis of interdisciplinary literature, integrating perspectives from computer science, public administration, law, and ethics. Its systematic methodology enhances the rigor and reliability of its findings. However, the authors acknowledged limitations stemming from the lack of empirical validation in real-world policy environments and from limited citizen participation in AI governance processes. Future research is recommended to test governance frameworks in practice and explore mechanisms for increasing public engagement.
Papadakis et al. (2024) examined the role of explainable and transparent artificial intelligence in public policymaking, focusing on how machine learning systems could support evidence-based decision-making while meeting regulatory and governance requirements. The study demonstrated that while AI improved the speed and analytical capacity of policymaking processes, its effectiveness in public-sector contexts depended on the ability to provide interpretable and transparent outputs. The findings highlighted that issues such as algorithmic bias, lack of explainability, and limited transparency created barriers to adoption, particularly in environments where decisions required justification to stakeholders and compliance with regulatory standards.
The study contributed to the literature by presenting and implementing a practical AI governance architecture that integrated the CRISP-DM process into a public-sector policy environment through the AI4PublicPolicy platform. The results showed that this framework enabled the acquisition and management of datasets, evaluation of multiple machine learning models, and generation of explainable outputs that supported policymaker interpretation. Additionally, the implementation of the QARMA framework demonstrated that human-readable rules could improve transparency by reducing reliance on opaque models, thereby enhancing the usability of AI systems in high-stakes policy contexts. These findings provided empirical support for the role of explainability in strengthening governance, transparency, and accountability in public-sector AI adoption.
The validation component further strengthened the study by demonstrating that the proposed approach was applied in real-world public-sector use cases involving local governments and legacy data systems. The results indicated that policymakers were able to evaluate model performance and utilize explainable outputs to inform decision-making processes. This practical application illustrated that explainable AI could be operationalized through cloud-based platforms, bridging the gap between theoretical governance principles and real-world implementation. However, the study also identified a key limitation: the absence of standardized criteria for evaluating the quality of explanations and the ongoing trade-off between model performance and interpretability.
These findings were directly relevant to the present study, as they provided empirical evidence that transparency and explainability influenced how practitioners interpreted and applied AI-driven outputs in governance contexts. Within the Technology Acceptance Model (TAM), the study’s results supported the relationship between perceived transparency and perceived usefulness, as interpretable outputs enhanced trust and decision-making confidence. Similarly, the integration of governance processes and explainable rules aligned with perceived compliance clarity by demonstrating how structured frameworks improved usability and implementation. The findings also supported the importance of perceived accountability, as explainable systems enabled clearer justification of decisions and reinforced institutional oversight. Collectively, the study reinforced the need for practice-based governance frameworks, such as the NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0), that operationalize transparency, accountability, and compliance in public-sector AI environments.
Wang and Blok (2025) critically examined the limitations of current AI ethics approaches, arguing that the shift from principle-based frameworks to practice-oriented methods such as Ethics-by-Design and Value-Sensitive Design remained largely confined to artifact-level concerns. The authors contended that these approaches often reflected techno-solutionist assumptions that failed to account for broader structural dynamics, including power asymmetries, institutional inequality, and socio-political complexity. Using an input–throughput–output analytic lens, they identified key constraints such as limited stakeholder participation, symbolic engagement, and assumptions of predictability that restricted the ability of existing governance models to effectively capture the full scope of socio-technical risks in public-sector environments where transparency, accountability, and regulatory compliance were critical.
Wang and Blok (2025) further argued that participation in AI governance processes was frequently constrained by organizational and economic factors, resulting in limited stakeholder influence over system design and deployment. These constraints, combined with the inherent complexity and nonlinearity of socio-technical systems, undermined the assumption that ethical outcomes could be fully anticipated or controlled through existing governance mechanisms. As a result, current models may have inadequately addressed systemic risks, particularly in government contexts that required robust accountability and transparency across multiple institutional levels.
To address these structural limitations, Wang and Blok (2025) proposed a multi-level framework that extended ethical analysis beyond technical systems to include organizational, socio-political, and ontological dimensions. This framework emphasized that AI-related risks were embedded within broader governance structures and societal contexts, requiring a more comprehensive approach to evaluation and oversight. By incorporating these layers, the framework provided a more holistic lens for identifying governance gaps that may not have been visible through traditional, design-focused approaches.
This perspective was directly relevant to the present study, as it highlighted how structural limitations in governance frameworks may have influenced practitioners’ perceptions of their effectiveness. Within the Technology Acceptance Model (TAM), these limitations may have reduced perceived usefulness when frameworks failed to account for real-world complexity and systemic conditions. Similarly, perceived transparency may have been weakened if governance mechanisms did not adequately capture underlying risks or power dynamics, while perceived accountability may have declined when responsibility was not clearly defined across institutional levels. Compliance clarity may also have been diminished when frameworks emphasized high-level ethical principles without providing actionable guidance for implementation.
The study identified a critical research gap in the lack of governance frameworks that integrated structural analysis with practical application. While existing literature acknowledged systemic challenges in AI ethics, it provided limited guidance on how to operationalize these insights within governance models. This gap was particularly significant in public-sector contexts, where agencies relied on frameworks such as the NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0) to guide responsible AI adoption. Without incorporating multi-level considerations, such frameworks may have been perceived as incomplete by practitioners responsible for implementation. By extending AI ethics beyond artifact-level design, Wang and Blok (2025) provided a theoretical foundation for advancing governance models that addressed both technical and structural challenges, supporting this study’s focus on how perceived transparency, accountability, and compliance clarity influenced the effectiveness of AI governance frameworks.
Elliott et al. (2024) examined how the rapid evolution of generative artificial intelligence (AI) disrupts traditional accountability structures in governance, particularly within public-sector and regulated environments. The authors argued that generative AI systems introduce a fundamentally different accountability paradigm, where responsibility is no longer linear or easily attributable to a single actor. Instead, accountability became “entangled” across multiple stakeholders, including developers, cloud providers, government agencies, data contributors, and end users. This diffusion of responsibility challenges existing governance models that rely on clear lines of authority, oversight, and liability.
The study emphasizes that generative AI systems, particularly those deployed through cloud-based platforms, operate within complex socio-technical ecosystems where decision-making authority is distributed. Unlike traditional information systems, generative AI continuously evolves through training, fine-tuning, and user interaction, making it difficult to determine who is accountable for outputs, errors, or unintended consequences. The authors highlight that this dynamic creates gaps in regulatory compliance, auditability, and oversight, as existing frameworks are not designed to accommodate adaptive, probabilistic systems. As a result, accountability must be reconceptualized as a shared and ongoing process rather than a static assignment of responsibility. A key contribution of the article is its identification of “accountability entanglement” as a defining characteristic of generative AI governance. This concept captures how accountability relationships are distributed across organizational, technical, and institutional layers, often spanning public and private sector boundaries. The authors argue that governance approaches must move beyond traditional hierarchical accountability models and instead incorporate networked and relational accountability structures. These structures should include mechanisms such as continuous monitoring, traceability of model outputs, shared governance agreements between stakeholders, and adaptive regulatory oversight.
The findings also underscore the limitations of current AI governance frameworks, which often emphasize transparency and explainability without fully addressing how responsibility is operationalized across stakeholders. While transparency may improve visibility into AI processes, it does not inherently resolve questions about who is responsible for decisions generated by complex AI systems. This reinforces the need for governance models that integrate accountability mechanisms throughout the AI lifecycle, including design, deployment, and post-deployment monitoring. The article identifies a critical research gap: the lack of empirical, practice-based frameworks to address accountability in distributed AI ecosystems. Specifically, there is limited understanding of how public-sector organizations can operationalize accountability when relying on third-party cloud providers and generative AI services. This gap is particularly relevant for government agencies, where statutory requirements demand clear accountability, yet AI systems introduce ambiguity in responsibility and control.
This study is highly relevant to the present research, as it directly supports the role of perceived accountability as a key determinant of governance effectiveness. Within the Technology Acceptance Model (TAM), the concept of accountability entanglement suggests that practitioners’ perceptions of accountability may be influenced not only by internal governance structures but also by their understanding of external dependencies and shared responsibilities. The findings reinforce the importance of developing governance frameworks that clearly define roles, responsibilities, and oversight mechanisms across the AI ecosystem, aligning with the study’s focus on evaluating the perceived effectiveness of the NIST AI RMF 1.0 in public-sector environments.
Straub et al. (2022) provided a comprehensive examination of artificial intelligence (AI) adoption in government, highlighting the need for standardized concepts, governance structures, and a unified framework to guide public-sector implementation. The authors argued that the rapid integration of AI into government operations had outpaced the development of coherent governance models, resulting in fragmented approaches to transparency, accountability, and regulatory compliance. This fragmentation created challenges for public-sector organizations attempting to align AI systems with statutory obligations and public expectations. A central contribution of the study is the development of a unified conceptual framework that integrates technical, organizational, and policy dimensions of AI governance. The authors emphasize that effective AI governance requires more than technical standards; it must also incorporate institutional accountability mechanisms, clear policy guidance, and cross-agency coordination. The framework highlights the importance of standardizing key terms, risk categories, and governance processes to ensure consistency in AI adoption across government entities. Without such standardization, agencies may interpret governance principles differently, leading to inconsistent implementation and reduced effectiveness of oversight mechanisms.
The study further identified transparency and accountability as foundational elements of trustworthy AI in government. Transparency was positioned as a mechanism for enabling oversight and public trust, while accountability ensured that decision-making processes remain subject to human control and institutional responsibility. However, the authors note that existing governance approaches often failed to operationalize these principles in a way that was actionable for practitioners. This gap between conceptual guidance and practical implementation limits government organizations’ ability to effectively manage AI risks and maintain compliance with regulatory frameworks.
Additionally, Straub et al. (2022) highlighted the importance of interoperability and shared standards across agencies and jurisdictions. As AI systems are increasingly deployed across interconnected government environments, the lack of unified standards can lead to inefficiencies, duplication of effort, and governance gaps. The authors advocate developing common frameworks that align technical implementation with policy objectives, enabling more effective coordination and oversight. The article identifies a significant research gap in the lack of empirically validated, practice-based governance frameworks for AI in government settings. While conceptual models and policy recommendations exist, there is limited research on how these frameworks are applied in real-world settings or on practitioners’ perceptions of their effectiveness. This gap is particularly relevant to local government contexts, where resources, expertise, and governance maturity may vary significantly.
This study directly supports the present research by reinforcing the importance of perceived transparency, accountability, and compliance clarity as critical factors influencing the effectiveness of AI governance frameworks. Within the Technology Acceptance Model (TAM), the unified framework proposed by Straub et al. (2022) aligns with perceived usefulness, as standardized governance structures can enhance oversight and decision-making effectiveness. At the same time, the emphasis on clear definitions and processes supports perceived ease of use, particularly regarding compliance clarity. These findings further justify empirically evaluating how government professionals perceive the effectiveness of frameworks such as the NIST AI RMF 1.0, particularly in environments characterized by fragmented governance and an increasing reliance on cloud-based AI systems.
de Fine Licht and de Fine Licht (2020) critically examined the role of transparency in artificial intelligence (AI) governance, particularly within the context of public decision-making. The authors argued that transparency was often positioned as a primary mechanism for ensuring accountability and legitimacy in AI-enabled systems; however, its effectiveness was contingent upon how transparency was conceptualized and operationalized. In public-sector environments, where decisions must be both explainable and justifiable to diverse stakeholders, the complexity of AI systems presents significant challenges to achieving meaningful transparency. The authors emphasized that simply increasing access to technical information does not necessarily improve understanding or accountability, particularly when stakeholders lack the expertise to interpret algorithmic processes.
A central contribution of the study is its distinction between different forms of transparency, including openness, explainability, and intelligibility. de Fine Licht and de Fine Licht (2020) highlight that transparency has to be tailored to its intended audience, as overly technical disclosures may fail to enhance accountability or trust. Instead, the authors advocate for “functional transparency,” which prioritizes the communication of relevant and understandable information that enables stakeholders to evaluate decisions and hold institutions accountable. This perspective challenges the assumption that more transparency inherently leads to better governance outcomes and underscores the need for carefully designed transparency mechanisms within AI systems.
The study further explored the relationship between transparency and accountability, arguing that transparency alone was insufficient to ensure responsible AI governance. Without accompanying accountability structures such as clear lines of responsibility, oversight mechanisms, and institutional enforcement transparency did not translate into meaningful governance outcomes. This limitation is particularly relevant in public-sector contexts, where decision-making authority must remain traceable and subject to democratic oversight. The authors suggest that effective governance requires the integration of transparency with broader institutional frameworks that support accountability and regulatory compliance.
Additionally, the article identifies key challenges in implementing transparency in AI systems, including trade-offs between transparency and other priorities such as privacy, security, and system performance. These competing demands highlight the need for balanced governance approaches that consider both technical and institutional constraints. The authors conclude that transparency should not be viewed as a standalone solution but rather as one component of a comprehensive governance strategy that incorporates accountability, oversight, and stakeholder engagement.
This study is directly relevant to the present research, as it reinforces the importance of perceived transparency and perceived accountability as interdependent factors influencing the effectiveness of AI governance frameworks. From a Technology Acceptance Model (TAM) perspective, the concept of functional transparency aligns closely with perceived usefulness, as clear and understandable explanations enhance stakeholders’ ability to interpret AI-driven decisions and build trust in system outputs. At the same time, the emphasis on making transparency accessible and interpretable supports perceived ease of use, particularly in the context of compliance clarity, where practitioners must understand governance requirements to effectively implement them (Venkatesh & Davis, 2000; Holden & Karsh, 2010). By highlighting the limitations of purely technical transparency and the need for actionable, user-centered explanations, this study provides a strong theoretical foundation for examining how transparency influences practitioners’ perceptions of governance framework effectiveness, particularly within frameworks such as the NIST Artificial Intelligence Risk Management Framework (AI RMF 1.0).
Wirtz et al. (2023) emphasized accountability as a critical dimension of AI governance, highlighting that responsibility for AI-driven decisions often became ambiguous within complex organizational environments. As AI systems are integrated into public-sector operations, organizations were required to establish clearly defined accountability structures that delineate roles, responsibilities, and oversight mechanisms to ensure that decision-making remains subject to human control. In the absence of such structures, authority may become diffused across technical systems and organizational actors, increasing the risk of governance gaps and complicating compliance and regulatory enforcement. This challenge is further intensified in cloud-based AI environments, where third-party providers contribute to system design, development, and operation, thereby introducing additional layers of complexity in assigning responsibility and maintaining effective oversight.
In addition to governance concerns, Wirtz et al. (2023) identified several organizational barriers to AI adoption, including limited technical expertise, insufficient data infrastructure, and a lack of standardized implementation frameworks. These constraints can hinder government agencies’ ability to operate AI governance principles effectively, leading to fragmented or inconsistent practices. The authors argue that addressing these barriers requires investment in institutional capacity, including workforce development, data governance strategies, and the establishment of clear regulatory guidelines.
The article also highlighted the importance of aligning AI governance with broader public administration principles, such as legality, transparency, and democratic accountability. The authors suggest that AI governance frameworks were required to be designed to support these principles while also accommodating the unique characteristics of AI systems, including their complexity, adaptability, and reliance on large-scale data. This alignment is essential for ensuring that AI adoption contributes to public value rather than undermining trust or accountability in government institutions (Wirtz et al., 2023).
A significant research gap identified in the study is the lack of empirical research on how AI governance frameworks are implemented and perceived within public-sector organizations. While conceptual models and policy recommendations are widely discussed, there is limited evidence on how practitioners interpret governance requirements or how these perceptions influence adoption and effectiveness. This gap is particularly relevant for local government contexts, where governance maturity and resource availability may vary significantly.
This study is highly relevant to the present research, as it reinforces the importance of perceived transparency, perceived accountability, and perceived compliance clarity as critical factors influencing the effectiveness of AI governance frameworks. Within the Technology Acceptance Model (TAM), transparency enhances perceived usefulness by improving understanding and trust, while accountability strengthens confidence in governance mechanisms and oversight processes (Venkatesh & Davis, 2000; Holden & Karsh, 2010). Additionally, the challenges of operationalizing governance principles align with the concept of compliance clarity, underscoring the need for frameworks that are both interpretable and actionable for public-sector practitioners. These findings support the study’s focus on evaluating the perceived effectiveness of the NIST AI Risk Management Framework (AI RMF 1.0) in guiding responsible AI adoption in local government settings.
The author Liu (2025) examined the foundational principles of artificial intelligence (AI) governance, with a particular focus on the ethical, regulatory, and operational considerations required for responsible AI adoption. The author argued that as AI systems became increasingly embedded in organizational and public-sector decision-making processes, governance frameworks needed to evolve to ensure alignment with ethical standards, regulatory requirements, and societal expectations. While AI technologies offer opportunities for efficiency and innovation, they also introduce risks related to bias, lack of transparency, and diminished accountability, which must be addressed through structured governance approaches (Liu, 2025).
A central contribution of the study is its emphasis on integrating ethical principles into governance frameworks. The author highlighted the core values such as fairness, transparency, accountability, and trustworthiness are essential for guiding AI system design and deployment. However, like broader AI governance literature, the study notes that these principles are often articulated at a high level and lack clear mechanisms for operationalization. This disconnect between ethical intent and practical implementation creates challenges for organizations attempting to translate governance principles into actionable processes (Liu, 2025).
The article further explored the role of regulatory and policy frameworks in shaping AI governance emphasized that governance mechanisms should incorporate monitoring, auditing, and continuous evaluation processes. It argues that effective governance requires alignment between organizational practices and external regulatory standards, including data protection laws, cybersecurity frameworks, and emerging AI-specific regulations. In public-sector contexts, this alignment is particularly critical, as agencies must demonstrate compliance with statutory requirements while maintaining transparency and public trust. The author emphasized that governance frameworks were expected to incorporate mechanisms for monitoring, auditing, and continuous evaluation to ensure that AI systems remain compliant and accountable over time (Lui, 2025).
Additionally, the study highlighted the importance of organizational readiness in implementing AI governance frameworks. Factors such as leadership commitment, technical expertise, data governance infrastructure, and cross-functional collaboration are identified as key enablers of effective governance. Without these capabilities, organizations may struggle to consistently implement governance frameworks, leading to fragmented or ineffective oversight. This reinforces the need for governance models that are not only conceptually robust but also practically applicable within real-world organizational environments (Liu, 2025).
The article identified a significant research gap in the lack of empirically validated frameworks that bridge the gap between ethical principles and operational governance practices. While the existing literature provided extensive theoretical guidance on responsible AI, there was limited research on how organizations implement these frameworks in practice or on practitioners’ perceptions of their usability and effectiveness. This gap was particularly relevant in public-sector environments, where governance requirements are complex, and resource constraints may limit implementation capacity.
This study is directly relevant to the present research, as it supported the importance of perceived transparency, perceived accountability, and perceived compliance clarity as key determinants of AI governance effectiveness. Within the Technology Acceptance Model (TAM), the clarity and operationalization of governance principles influence perceived ease of use, while transparency and accountability enhance perceived usefulness and trust (Venkatesh & Davis, 2000; Holden & Karsh, 2010). The findings further reinforced the need for governance frameworks, such as the NIST AI Risk Management Framework (AI RMF 1.0), to provide clear, actionable guidance that can be effectively implemented by public-sector practitioners.
Mökander et al. (2022) examined the practical challenges of implementing responsible artificial intelligence (AI) governance through ethics-based auditing of automated decision-making systems. The authors argued that while ethical principles were widely promoted, organizations often struggled to translate these principles into operational practices. Ethics-based auditing was presented as a mechanism to bridge this gap by systematically evaluating AI systems against ethical standards; however, the authors emphasize that such audits are not a comprehensive solution and must be integrated within broader governance structures.
A central contribution of the study is its conceptualization of ethics-based auditing as part of a larger socio-technical governance process. The authors highlight that auditing AI systems involves assessing not only technical components, such as algorithms and datasets, but also organizational practices, decision-making processes, and institutional accountability mechanisms (Mökander et al., 2022). This broader perspective reinforces the idea that AI governance extends beyond technical compliance to include organizational and regulatory dimensions, particularly in public-sector environments where accountability requirements are more stringent. The study further explored transparency and accountability as key pillars of effective AI governance. Transparency was framed as the ability to provide meaningful insight into how automated decisions are generated, including the documentation of data sources, model design, and decision logic. However, the authors noted that transparency alone is insufficient without corresponding accountability mechanisms that define responsibility for AI outcomes and enable oversight and redress (Mökander et al., 2022). This finding aligns with broader governance literature, which emphasizes that accountability structures must be clearly defined and operationalized to ensure trustworthy AI deployment.
In addition, Mökander et al. (2022) identified several limitations of ethics-based auditing, including reliance on organizational commitment, difficulty in standardizing audit processes, and challenges in capturing the full complexity of AI systems. Audits may also be constrained by limited access to proprietary systems or insufficient technical expertise, particularly in organizations that depend on third-party vendors or cloud-based AI services. These limitations highlight the need for complementary governance mechanisms, such as continuous monitoring, regulatory oversight, and standardized frameworks, to ensure comprehensive AI governance.
The article identified a significant research gap in the lack of empirical evidence on the effectiveness of ethics-based auditing in real-world settings. While auditing is increasingly promoted as a best practice for responsible AI, there is limited research on how organizations implement these processes or on practitioners’ perceptions of their usefulness and usability (Mökander et al., 2022). This gap is particularly relevant for public-sector organizations, where governance frameworks must be both rigorous and practical to support compliance and accountability requirements.
This study is directly relevant to the present research, as it reinforced the importance of perceived transparency, perceived accountability, and perceived compliance clarity in shaping the effectiveness of AI governance frameworks. Within the Technology Acceptance Model (TAM), transparency contributes to perceived usefulness by enhancing understanding and trust, while accountability strengthens confidence in governance mechanisms (Venkatesh & Davis, 2000; Holden & Karsh, 2010). Furthermore, the challenges associated with operationalizing ethics-based auditing align with the concept of compliance clarity, underscoring the need for governance frameworks, such as the NIST AI Risk Management Framework (AI RMF 1.0), that provide clear, actionable guidance that can be effectively implemented by public-sector practitioners.
Transparency and explainability have emerged as foundational principles of trustworthy artificial intelligence (AI) governance because they enable stakeholders to understand how AI systems generate recommendations, support organizational accountability, and foster public trust. As AI becomes increasingly integrated into public-sector operations, government agencies must demonstrate that automated decisions are understandable, ethically justified, and aligned with regulatory expectations. Recent scholarship suggests that transparency extends beyond technical disclosure and encompasses governance practices that clearly communicate decision-making processes, responsibilities, and risk management activities (Criado, 2025; Liu, 2025; Porumbescu et al., 2025).
Liu (2025) argued that transparency should be viewed as a multidimensional governance principle that integrates explainability, fairness, accountability, and ethics throughout the AI lifecycle rather than as a technical characteristic of machine learning models. The author emphasized that AI systems frequently operate as “black boxes,” limiting users’ ability to understand how decisions are generated and reducing confidence in AI-supported outcomes. To address these concerns, Liu proposed that governance frameworks incorporate explainable AI techniques, ethics-by-design principles, and multi-stakeholder governance models that improve both institutional accountability and public trust. The study further suggested that meaningful transparency requires organizations to provide understandable explanations of AI decisions while simultaneously establishing governance structures that ensure oversight and responsible implementation.
Recent research has expanded transparency beyond algorithmic explainability by examining its relationship with organizational acceptance and public value. Park and Jo (2025) suggested that regulatory perceptions influence how public-sector employees evaluate AI technologies and that governance mechanisms supporting transparency can strengthen confidence in AI-enabled decision-making. Similarly, Criado (2025) emphasized that successful AI adoption within public administration depends not only on technological capability but also on governance structures that communicate risks, responsibilities, and organizational expectations in ways that practitioners can readily understand. Collectively, these studies indicate that transparency contributes to greater organizational confidence by reducing uncertainty surrounding AI-supported decisions and facilitating informed governance practices.
Transparency also extends beyond organizational decision-making to broader institutional governance. Luitse (2024) argued that increasing dependence on hyperscale cloud providers creates new governance challenges by limiting visibility into AI infrastructures and reducing organizational control over data, models, and operational processes. As governments increasingly procure AI capabilities through commercial cloud providers, transparency becomes essential for maintaining accountability, demonstrating regulatory compliance, and preserving digital sovereignty. These findings suggest that transparency should encompass not only algorithmic explainability but also visibility into the broader governance ecosystem supporting AI deployment.
The importance of transparency is further reinforced by recent research on public trust and responsible AI governance. Vatamanu et al. (2025) emphasized that trustworthy AI requires governance mechanisms that integrate transparency, ethical oversight, and stakeholder engagement throughout the AI lifecycle. Likewise, Porumbescu et al. (2025) argued that public-sector organizations implementing generative AI must establish transparent governance practices that communicate AI capabilities, limitations, and oversight responsibilities to maintain public confidence. Together, these studies suggest that transparency is increasingly recognized as both a technical and institutional requirement for responsible AI adoption within government organizations.
Although the literature consistently identifies transparency and explainability as essential characteristics of trustworthy AI governance, most studies remain conceptual, policy-oriented, or focused on organizational best practices. Limited empirical evidence exists regarding how local government professionals perceive transparency when evaluating AI governance frameworks or whether those perceptions influence assessments of governance effectiveness. Consequently, additional research is needed to examine whether perceived transparency contributes to practitioners’ perceptions of the effectiveness of the National Institute of Standards and Technology Artificial Intelligence Risk Management Framework (NIST AI RMF 1.0) within local government organizations.
This chapter synthesized the existing literature on artificial intelligence (AI) governance within public-sector organizations, with a specific focus on perceived transparency, perceived accountability, and perceived compliance clarity as key determinants of governance effectiveness. The review highlighted that while AI adoption continues to accelerate across government environments, governance frameworks have not evolved at the same pace, resulting in fragmented oversight, inconsistent implementation, and challenges in maintaining regulatory compliance and public trust (Straub et al., 2022; Wirtz et al., 2023). Across the literature, transparency and accountability consistently emerged as foundational principles for trustworthy AI; however, studies emphasized that these constructs are often defined conceptually rather than operationalized in ways that are actionable for practitioners.
A recurring theme throughout the literature is the gap between high-level ethical and governance principles and their practical implementation. Research on compliance clarity demonstrated that public-sector practitioners require frameworks that provide clear, interpretable, and actionable guidance to effectively align AI systems with statutory and regulatory requirements (Jansen et al., 2023; Hussain et al., 2024). Similarly, studies on accountability revealed increasing complexity in AI ecosystems, particularly with the rise of cloud-based and generative AI systems, where responsibility is distributed across multiple stakeholders, creating challenges for oversight and enforcement (Elliott et al., 2024). These findings are further supported by research on ethics-based auditing and explainable AI, which indicate that transparency alone is insufficient without clearly defined accountability structures and continuous governance mechanisms (Mökander et al., 2022; Papadakis et al., 2024).
The literature also emphasized the importance of institutional readiness and governance standardization in supporting effective AI adoption. Public-sector organizations often face barriers such as limited technical expertise, lack of unified governance frameworks, and difficulties integrating ethical principles into operational processes (Wirtz et al., 2023; Straub et al., 2022). Additionally, emerging research highlighted structural limitations in AI governance, including power asymmetries, socio-technical complexity, and the inability of existing frameworks to address multi-level governance challenges (Wang & Blok, 2025). These limitations suggest that current governance models may not fully capture the complexity of AI deployment in public-sector environments.
A critical research gap identified across the literature is the lack of empirical, practitioner-focused studies evaluating how government professionals perceive the effectiveness of AI governance frameworks. While conceptual models and policy recommendations are well established, there is limited evidence on how these frameworks are interpreted, applied, and evaluated in real-world local government settings. This gap is particularly significant given the increasing reliance on cloud-based AI systems and the need for governance frameworks that are both comprehensive and practical.
The Technology Acceptance Model (TAM) provides relevant theoretical lens for addressing this gap by linking practitioner perceptions to framework effectiveness. Specifically, perceived transparency and accountability influence perceived usefulness, while compliance clarity aligns with perceived ease of use, collectively shaping the perceived effectiveness and adoption of governance frameworks (Venkatesh & Davis, 2000; Holden & Karsh, 2010). Building on these insights, this study advances the literature by empirically examining the relationship between these constructs and the perceived effectiveness of the NIST AI Risk Management Framework (AI RMF 1.0) within local government organizations. By addressing the identified research gap, this study contributes to both theory and practice by providing evidence-based insights into how AI governance frameworks can be improved to support transparent, accountable, and compliant AI adoption in the public sector.
Bannister, F., & Connolly, R. (2023). The future of digital government: The role of AI in public administration. Information Polity, 28(2), 123–136. https://doi.org/10.3233/IP-220071
Creswell, J. W., & Creswell, J. D. (2018). Research design: Qualitative, quantitative, and mixed methods approaches (5th ed.). SAGE Publications. https://us.sagepub.com/en-us/nam/research-design/book255675
Davis, F. D. (1986). A technology acceptance model for empirically testing new end-user information systems: Theory and results (Doctoral dissertation, Massachusetts Institute of Technology). MIT Libraries. https://hdl.handle.net/1721.1/15192
Davis, F. D. (1989). Perceived usefulness, perceived ease of use, and user acceptance of information technology. Journal of Risk and Uncertainty, 18(3), 321–325. https://doi.org/10.1023/a:1011156710779
de Fine Licht, J., & de Fine Licht, K. (2023). Artificial intelligence, transparency, and public decision-making. AI & Society, 38(1), 1–12. https://doi.org/10.1007/s00146-020-00960-w
Dwivedi, Y. K., Hughes, L., Ismagilova, E., Aarts, G., Coombs, C., Crick, T., Duan, Y., Dwivedi, R., Edwards, J., Eirug, A., Galanos, V., Ilavarasan, P. V., Janssen, M., Jones, P., Kar, A. K., Kizgin, H., Kronemann, B., Lal, B., Lucini, B., & Williams, M. D. (2023). So what if ChatGPT wrote it? Multidisciplinary perspectives on opportunities, challenges, and implications of generative AI. International Journal of Information Management, 71, 102642. https://doi.org/10.1016/j.ijinfomgt.2023.102642
Eisenberg, I. W., Gamboa, L., & Sherman, E. (2025, March 11). The unified control framework: Establishing a common foundation for enterprise AI governance, risk management and regulatory compliance. arXiv. https://arxiv.org/abs/2503.05937
Field, A. (2018). Discovering statistics using IBM SPSS statistics (5th ed.). SAGE Publications. https://us.sagepub.com/en-us/nam/discovering-statistics-using-ibm-spss-statistics/book246128
Folorunso, A., Adewa, A., Babalola, O., & Nwatu, C. E. (2024). A governance framework model for cloud computing: Role of AI, security, compliance, and management. World Journal of Advanced Research and Reviews, 24(2), 1969–1982. https://doi.org/10.30574/wjarr.2024.24.2.3513
Holden, R. J., & Karsh, B.-T. (2010). The Technology Acceptance Model: Its past and its future in health care. Journal of Biomedical Informatics, 43(1), 159–172. https://doi.org/10.1016/j.jbi.2009.07.002
Jansen, B., Kadenko, N., Broeders, D., Michel van Eeten, Borgolte, K., & Fiebig, T. (2023). Pushing boundaries: An empirical view on the digital sovereignty of six governments in the midst of geopolitical tensions. Government Information Quarterly, 40(4), 101862. https://doi.org/10.1016/j.giq.2023.101862
Janssen, M., Brous, P., Estevez, E., Barbosa, L., & Janowski, T. (2023). Data governance: Organizing data for trustworthy artificial intelligence. Government Information Quarterly, 40(1), 101493. https://doi.org/10.1016/j.giq.2020.101493
Kim, D., Lee, H., Lee, S., & Eom, S.-J. (2024). Conditions for AI systems adoption in public sector: From an accountability perspective. https://doi.org/10.1145/3657054.3657060
Knowles, B., Richards, J. T., & Kroeger, F. (2022). The many facets of trust in AI: Formalizing the relation between trust and fairness, accountability, and transparency. arXiv. https://doi.org/10.48550/arXiv.2208.00681
Kouziokas, G. N. (2023). Artificial intelligence in public administration: A review of governance and ethical challenges. Technological Forecasting and Social Change, 187, 122190. https://doi.org/10.1016/j.techfore.2022.122190
Li, H., & Sun, Z. (2025). Is Algorithmic Accessibility Sufficient? The Pivotal Role of Accessibility and Accountability in Shaping Trust in Automated Decision‐Making. Governance, 38(4). https://doi.org/10.1111/gove.70067
Liu, E. (2025). Building Trustworthy AI: Transparency, Fairness, and Governance in the Digital Age. 1(1), 3–5. https://doi.org/10.63802/afs.v1.i1.80
Lothery, E. J. (2024). Transformative governance: Integrating generative artificial intelligence in state and local government operations (Doctoral dissertation, Bowling Green State University). ScholarWorks @ BGSU. https://scholarworks.bgsu.edu/td/1874/
Luitse, D. (2024). Platform power in AI: The evolution of cloud infrastructures in the political economy of artificial intelligence. Internet Policy Review, 13(2). https://doi.org/10.14763/2024.2.1768
Nathan, C., & Hyams, K. (2021). Global policymakers and catastrophic risk. Policy Sciences, 55(1), 3–21. https://doi.org/10.1007/s11077-021-09444-0
Papadakis, T., Christou, I. T., Charalampos Ipektsidis, Soldatos, J., & Amicone, A. (2024). Explainable and transparent artificial intelligence for public policymaking. Data & Policy, 6. https://doi.org/10.1017/dap.2024.3
Park, E., & Jo, G. (2025). AI Acceptance Model for Public Value Creation: Regulatory Perception with the Moderating Effects of Social Expectancy Disconfirmation. Hyeondae Sahoe Wa Haengjeong, 35(2), 75–106. https://doi.org/10.26847/mspa.2025.35.2.75
Pasupuleti, M. K. (2025). Governance Frameworks for Ethical AI Deployment in Public Sector Services. 05(05), 561–573. https://doi.org/10.62311/nesx/rphcr17
Robles, P. (2023). Developing an effective artificial intelligence governance framework for public organizations (Doctoral dissertation, Pennsylvania State University). Pennsylvania State University ETDA. https://etda.libraries.psu.edu/files/final_submissions/28783
Romzek, B. S., & Ingraham, P. W. (2000). Cross pressures of accountability: Initiative, command, and failure in the Ron Brown plane crash. Public Administration Review, 60(3), 240–253. https://doi.org/10.1111/0033-3352.00085
Schmitt, M. (2024). Accountability in AI governance: Bridging the gap between policy and implementation. Policy & Internet. Advance online publication. https://doi.org/10.1002/poi3.345
Stanisław Galij, Pawlak, G., & Grzyb, S. (2024). Modeling Data Sovereignty in Public Cloud—A Comparison of Existing Solutions. Applied Sciences, 14(23), 10803–10803. https://doi.org/10.3390/app142310803
Straub, V. J., Morgan, D., Bright, J., & Margetts, H. (2022). Artificial intelligence in government: Concepts, standards, and a unified framework. arXiv. https://doi.org/10.48550/arXiv.2210.17218
U.S. Government Accountability Office. (2023). Artificial intelligence: Key practices to help ensure accountability in federal use (GAO-23-105747). https://www.gao.gov/products/gao-23-105747
Zuiderwijk, A., Chen, Y.-C., & Salem, F. (2022). Implications of the use of artificial intelligence in public governance: A systematic literature review and a research agenda. Government Information Quarterly, 39(3), 101577. https://doi.org/10.1016/j.giq.2021.101577
Figure 1
G*Power Sample Size
The SME asked for you to change this word and you haven’t. You need to change it to Proposal.
Resolved
Please review the DSE Template for Chapter 2. You have some headings that appear to be at the wrong level.
Resolved
Resolved
Resolved
Your references are starting indented and the shouldn’t. I’ve corrected the first reference so you have an example. Fix these.
Evan John