Agentic AI in Newsrooms: Towards a multi-dimensional framework for evaluating trust, editorial accountability, and workflow quality

 Corresponding author: Samson Agbaeze Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution License 4.0. Agentic AI in Newsrooms: Towards a multi-dimensional framework for evaluating trust, editorial accountability, and workflow quality Samson Emeka Agbaeze 1, *, Vivian Claire Okeke 2, Ebiere Precious Phillips 3, Amara Lucy Jacobs 4 and Samuel Tobi Oluwakoya 5 1 Digital Transformation, Business Administration, Nexford University, Washington, District of Columbia, US. 2 Media and Communication, Afe Babalola University, Ado-Ekiti, Ekiti, Nigeria. 3 Business Administration, Nexford University, Washington, District of Columbia, US. 4 Mass Communication, University of Calabar, Calabar, Nigeria. 5 Computer Science, Afe Babalola University, Ado-Ekiti, Ekiti, Nigeria. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 Publication history: Received on 30 September 2025; revised on 08 November 2025; accepted on 12 November 2025 Article DOI: https://doi.org/10.30574/wjarr.2025.28.2.3766 Abstract As artificial intelligence (AI) systems evolve from assistive to agentic capable of autonomous planning, decision-making, and content generation existing evaluation frameworks struggle to capture their broader organizational and ethical implications. Most assessments of newsroom AI focus narrowly on technical accuracy or efficiency, overlooking how such systems reshape trust, governance, and human collaboration. This study conducts a systematic literature review of 46 peer-reviewed and institutional sources (2015–2025) to examine how AI performance in journalism can be evaluated more holistically. Drawing from Information Systems Success Theory, Socio-Technical Systems Theory, Accountability Theory, and Trust Theory, the paper proposes a Four-Dimensional (4D) Evaluation Framework encompassing Technical Quality, Human–Organizational Alignment, Ethical–Governance Responsibility, and Trust–Value Impact. The framework reconceptualizes AI success as a socio-technical equilibrium where technological capacity, ethical integrity, and collaborative trust co-evolve. It contributes to the emerging field of Responsible AI in journalism by offering a multi-dimensional structure for evaluating agentic AI systems that balances innovation with accountability and public value. Keywords: Agentic AI; Journalism; Information Systems Evaluation; Trust; Ethics; Responsible AI; Socio-Technical Systems 1. Introduction The rapid infusion of artificial intelligence (AI) into global newsrooms marks one of the most profound transformations in the communication industry since the rise of digital journalism. Across continents, media organizations now deploy algorithmic systems to automate content generation, personalize audience experiences, detect misinformation, and optimize newsroom workflows. While these technologies promise efficiency and scale, their increasing autonomy raises deeper questions about trust, editorial accountability, and the overall quality of journalistic work. In recent years, agentic AI systems AI tools capable of planning, decision-making, and acting with minimal human oversight have emerged as new “actors” within the newsroom ecosystem (Baird and Maruping, 2021; Dörr, 2016). Yet, despite their growing presence, there is still no coherent framework to evaluate their impact on journalism’s institutional integrity and democratic purpose. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1062 Existing research reflects fragmented scholarly attention. Studies in the early wave of automated journalism (2014– 2020) focused on the technical accuracy of machine-generated stories and audience perceptions of credibility (Hansen et al., 2017; Dörr, 2016). More recent work, particularly from the Reuters Institute, FIAT/IFTA, and Tow Center for Digital Journalism, explores newsroom-level adoption and ethical guidelines for AI-assisted reporting (Beckett, 2019; Diakopoulos, 2019). However, these studies are largely descriptive, examining what tools are used or how journalists feel about them. They rarely interrogate how agentic AI systems influence the deeper socio-technical relationships that define journalistic quality, editorial responsibility, and public trust. Consequently, while the capabilities of AI in journalism are increasingly well understood, their evaluation remains narrowly defined by technical performance metrics rather than social or ethical outcomes. This lack of holistic evaluation presents a significant research gap. Traditional Information Systems (IS) evaluation models—such as the DeLone and McLean IS Success Model—offer multidimensional perspectives on system success, including user satisfaction and organizational impact (DeLone and McLean, 2016). Yet, their application in journalism remains limited. Similarly, theories of accountability and transparency in media systems provide conceptual lenses for assessing editorial responsibility (Plaisance, 2015; Karlsson and Clerwall, 2018), but these are seldom integrated with IS or socio-technical frameworks. The absence of interdisciplinary synthesis has created an imbalance: AI systems are judged by efficiency, not by their implications for ethical governance or journalistic values. This study seeks to bridge that divide. The aim of this research is therefore to develop a multi-dimensional, IS-grounded evaluation framework for assessing the role and impact of agentic AI in newsrooms. Specifically, it examines how such systems affect three interrelated domains audience trust, editorial accountability, and workflow quality—drawing from both Information Systems theory and Communication research. By synthesizing peer-reviewed studies, institutional reports, and conceptual works from 2015 to 2025, this paper proposes an integrated evaluative model that aligns technological, human, and organizational dimensions of newsroom AI. This study contributes in three major ways. First, it provides a systematic synthesis of how AI has been conceptualized and applied within newsroom contexts, identifying trends, tensions, and theoretical blind spots. Second, it extends IS and communication theories to a new frontier agentic AI in journalism by introducing a framework that treats AI as both a technological artifact and an organizational actor. Finally, it offers a practical roadmap for news organizations, regulators, and scholars seeking to design or audit responsible AI systems that reinforce, rather than erode, the ethical foundations of journalism. In doing so, this paper aligns with global debates on algorithmic accountability, data governance, and trust restoration in digital media ecosystems. 2. Theoretical Background and Conceptual Foundations Understanding the impact of agentic AI in newsrooms requires a theoretical base that captures both its technical functions and its organizational consequences. Journalism, as a socio-technical profession, has long been shaped by the interaction between human judgment, technological tools, and institutional norms. As AI becomes an increasingly autonomous collaborator in news production, these interactions intensify and transform. To assess this transformation holistically, the study draws on four theoretical foundations that, together, illuminate how AI reshapes newsroom trust, accountability, and workflow: the Information Systems (IS) Success Model, Socio-Technical Systems (STS) Theory, Accountability and Transparency Theory, and Trust Theory. 2.1. Information Systems Success Model The DeLone and McLean IS Success Model remains one of the most influential frameworks for evaluating the performance of information systems across industries. Originally introduced in 1992 and updated in 2003 and 2016, the model proposes six interdependent dimensions of system success: system quality, information quality, service quality, user satisfaction, intention to use, and net benefits (DeLone and McLean, 2016). In newsroom contexts, these dimensions can translate into critical measures such as the reliability of AI-generated outputs, perceived usefulness by journalists, and the broader organizational value derived from AI-supported editorial workflows. However, existing studies of AI in journalism have rarely extended the DeLone and McLean model to include ethical and human-centered outcomes. Technical measures like accuracy or error rate dominate evaluation, while the effects on editorial independence, fairness, and public accountability remain understudied (Diakopoulos, 2019; Hansen, 2017). By adapting the IS Success Model to the newsroom, this study expands the definition of “success” to incorporate socioethical dimensions, arguing that an AI system cannot be deemed successful merely because it performs efficiently—it must also uphold journalistic integrity and societal trust. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1063 2.2. Socio-Technical Systems (STS) Theory While the IS Success Model focuses on system outcomes, Socio-Technical Systems (STS) Theory explains how those outcomes emerge from the interaction between social and technical subsystems within an organization (Trist and Bamforth, 1951; Mumford, 2006). In the newsroom, this means that the effectiveness of AI tools depends not only on their technical capacity but also on how journalists, editors, and managers integrate them into professional routines and ethical decision-making. STS theory thus provides a lens for examining how human agency coexists—and sometimes conflicts—with machine agency. Applying STS to journalism underscores that technology adoption is never neutral. Studies show that AI tools can redistribute decision-making power in subtle ways: algorithms may prioritize certain story types, suggest headlines that optimize engagement, or automatically flag content for ethical review (Beckett, 2019; Dörr, 2016). These affordances can increase productivity but may also shift editorial control away from journalists toward opaque algorithmic processes. Therefore, evaluating AI systems in newsrooms must consider both the technical performance of the system and the organizational adaptation it triggers. STS theory aligns perfectly with this study’s goal of building a multi-dimensional evaluation model that recognizes both human and machine agency. 2.3. Accountability and Transparency Theory Journalism’s moral legitimacy rests on its ability to be accountable to the public. Accountability Theory in media ethics emphasizes mechanisms such as editorial oversight, corrections, and transparency disclosures that allow audiences to evaluate journalistic credibility (Plaisance, 2015; Karlsson and Clerwall, 2018). When AI participates in editorial decisions, these accountability structures face new challenges. Who is responsible when an AI-generated article contains bias or misinformation the developer, the editor, or the algorithm itself? Contemporary research suggests that transparency the act of explaining how AI systems work is critical to sustaining public confidence (Vos and Craft, 2017). Yet transparency in AI journalism remains superficial. Many organizations disclose that AI tools are used but rarely clarify how data is processed or what editorial safeguards are applied (Jamil, 2023; Fernández and Serrano, 2025). Accountability theory thus complements STS by grounding evaluation in ethical responsibility: it demands that the “black box” of newsroom AI be opened to public scrutiny. Incorporating these principles into AI evaluation frameworks ensures that efficiency gains do not come at the cost of editorial accountability or media credibility. 2.4. Trust Theory Trust has always been central to journalism’s social contract, and the arrival of AI amplifies this dependency. Trust Theory, particularly the integrative model proposed by Mayer, Davis, and Schoorman (1995), defines trust as the willingness to be vulnerable to another entity’s actions based on perceptions of ability, benevolence, and integrity. In AI-mediated journalism, trust operates at two levels: (1) internal trust, between journalists and the AI tools they use, and (2) external trust, between audiences and AI-assisted outputs. Recent studies highlight growing skepticism toward AI-generated content, with audiences questioning its authenticity and ethical grounding (Cools and Koliska, 2024; Wölker and Powell, 2021). At the same time, journalists themselves exhibit ambivalence—valuing AI’s efficiency but doubting its judgment (Jamil, 2023). By applying trust theory, this study situates trust not as a passive outcome but as a relational process involving design transparency, ethical governance, and consistent performance. When integrated with accountability and socio-technical perspectives, trust becomes both an evaluative dimension and an indicator of overall newsroom health. 2.5. Integrating Theories into a Unified Lens Each of these theoretical perspectives captures a crucial piece of the newsroom-AI puzzle. The IS Success Model explains what constitutes success; STS Theory describes how human and machine systems interact; Accountability Theory defines why ethical responsibility matters; and Trust Theory clarifies how legitimacy is sustained. Together, they provide a foundation for a multi-dimensional evaluation framework that moves beyond technical assessment to encompass human, ethical, and institutional dimensions. Integrating these perspectives allows the development of a 4D Evaluation Model for Agentic AI in Newsrooms, which this study later proposes. The four dimensions—technical, organizational, ethical, and trust-based—reflect the interdependent forces shaping AI’s role in journalism today. This synthesis answers the core research question driving this paper: How can agentic AI in newsrooms be evaluated in a way that balances technological performance with ethical responsibility and audience trust? World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1064 By grounding the discussion in well-established theories, this study ensures that the framework is not only conceptually rigorous but also adaptable for empirical validation in future research. It thereby contributes to an emerging scholarly consensus that journalism’s AI transformation must be studied as a complex, socio-technical evolution rather than a purely technological disruption. 3. Methodology: systematic literature review approach This study employs a systematic literature review (SLR) design to synthesize academic and professional evidence on the evaluation of agentic AI in newsrooms. The approach follows the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) protocol, ensuring methodological transparency and reproducibility (Page et al., 2021). An SLR is appropriate because research on newsroom-level AI is dispersed across communication, informationsystems, and media-ethics disciplines; a structured review enables integration and theory building without collecting new primary data. 3.1. Review Design and Objectives The review pursued three objectives • Identify peer-reviewed and institutional literature (2015–2025) that examines the adoption, governance, and ethical evaluation of AI in journalism. • Extract and categorize findings along the three dimensions highlighted in this study—trust, editorial accountability, and workflow quality. • Synthesize insights into an integrated, IS-grounded framework for evaluating agentic AI in newsrooms. The ten-year window (2015–2025) captures the evolution from early automated-journalism tools to contemporary generative and agentic systems. The review combined academic and industry sources to ensure both theoretical depth and applied relevance. 3.2. Data Sources and Search Strategy Searches were conducted in Scopus, Web of Science, Google Scholar, and subject-specific databases such as Communication and Mass Media Complete and ACM Digital Library. To incorporate grey literature, institutional repositories from the Reuters Institute for the Study of Journalism, Tow Center for Digital Journalism, UNESCO, and Knight Foundation were included. The Boolean search string combined key concepts and synonyms (“artificial intelligence” OR “algorithmic journalism” OR “automated journalism” OR “agentic AI”) AND (“newsroom” OR “journalism” OR “media production”) AND (“trust” OR “accountability” OR “workflow” OR “ethics” OR “evaluation”). Each database search was limited to English-language publications between 2015 and 2025. Reference lists of retrieved papers were manually scanned to capture additional relevant studies (the snowball technique). 3.3. Inclusion and Exclusion Criteria 3.3.1. Inclusion criteria • Peer-reviewed journal articles, conference papers, books, or recognized institutional reports. • Explicit focus on AI systems in journalism, newsroom management, or editorial workflows. • Discussion of at least one target dimension: trust, accountability, or workflow quality. 3.3.2. Exclusion criteria • Purely technical studies with no organizational or ethical dimension (e.g., model-training papers). • Opinion pieces or short news items without methodological grounding. • Duplicates or papers unavailable in full text. After applying these criteria, 214 initial records were identified; 162 remained after duplicates were removed. Screening of titles and abstracts yielded 73 eligible sources, and full-text review reduced these to 46 studies that directly informed the thematic synthesis. (A PRISMA flow diagram will visualize this process in Section 4.) World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1065 3.4. Data Extraction and Coding Process A coding matrix was developed in Excel, capturing for each study: • Author(s), year, country, journal/source. • Method type (quantitative, qualitative, conceptual, mixed). • AI application domain (content creation, editing, distribution, ethics). • Key findings and implications for trust, accountability, workflow quality. Codes were iteratively refined following inductive thematic analysis (Braun and Clarke, 2006). Each study could be coded under multiple dimensions to capture intersectionality for instance, a paper on AI-assisted headline generation could inform both workflow quality and trust. To enhance reliability, a second coder independently reviewed a 20 percent sample of the dataset, achieving Cohen’s κ = 0.84, indicating substantial agreement (Landis and Koch, 1977). 3.5. Synthesis and Analytical Strategy The final synthesis proceeded in two stages: • Descriptive mapping quantified publication trends by year, region, and method, providing a panoramic view of how the field has evolved. • Thematic integration distilled cross-cutting insights into the three analytical categories guiding this research. Themes were then conceptually aligned with the four theoretical pillars discussed earlier (IS Success, STS, Accountability, Trust). This dual-stage analysis ensures both breadth and interpretive depth, allowing the review to transition seamlessly from data patterns to theoretical generalization. 3.6. Validity, Reliability, and Limitations To maintain transparency, all search terms, databases, and coding decisions were documented in an audit trail. Triangulation between academic and institutional sources mitigated disciplinary bias. Nonetheless, the review is limited by the predominance of Western literature; studies from Africa, Asia, and Latin America remain underrepresented, reflecting a global imbalance in AI-journalism research (Jamil, 2023). Future empirical work should address this gap through cross-regional comparisons. Despite these constraints, the systematic approach provides a rigorous foundation for developing the 4D Evaluation Framework presented in the next section. By linking methodological precision with theoretical synthesis, the study strengthens the credibility of its conceptual contribution and supports its replicability for future research. 4. Results and Thematic Synthesis The systematic review yielded 46 eligible studies that directly examined AI’s role in news production between 2015 and 2025. These included 32 peer-reviewed journal articles, 6 conference papers, and 8 professional or institutional reports. Following PRISMA procedures, the review process ensured transparency and replicability. 4.1. PRISMA Flow of Study Selection The search and screening process followed the PRISMA 2020 guidelines (Page et al., 2021). The complete flow is summarized below and visually represented in Figure 1. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1066 Figure 1 PRISMA Flow Diagram for Study Selection The PRISMA diagram clarifies the filtering process that led from 214 initial records to 46 final sources forming the synthesis base. The inclusion of both peer-reviewed and professional reports allowed for balanced coverage of theoretical development and newsroom practice. This process is also supported by the Screening and Coding Matrix (Appendix B), which lists all 46 studies, their metadata (author, year, focus, and method), and thematic assignments. 4.2. Descriptive Overview of the Literature Table 1 provides a descriptive snapshot of the reviewed studies. A clear pattern emerges: research output accelerates sharply after 2020, coinciding with the availability of large-language models and newsroom experimentation with generative AI. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1067 Table 1 Overview of reviewed studies (2015–2025) Category Count % of total Notes Publication type Journal articles 32 69.6 Peer-reviewed (Digital Journalism, Journalism Studies, etc.) Reports / white papers 8 17.4 Reuters Institute, Tow Center, UNESCO Conference papers 6 13.0 ICA, IAMCR, AMCIS presentations Geographic focus Europe and North America 28 60.9 Major news organizations (BBC, AP, NYT) Asia and Middle East 10 21.7 Emerging AI use in local newsrooms Africa and Latin America 8 17.4 Sparse but growing interest (Jamil, 2023) Dominant method Qualitative / case study 22 47.8 Interviews, ethnography, content analysis Quantitative / survey 11 23.9 Audience trust, newsroom readiness Conceptual / theoretical 13 28.3 Frameworks, ethical reflections (Sources compiled from Appendix B) The dataset shows that while Western contexts dominate, cross-regional interest is increasing. Conceptual and ethical studies now represent nearly one-third of publications, signaling the field’s shift from description to evaluation—an essential foundation for this paper’s proposed framework. 4.3. Theme 1: Trust in AI-Assisted Journalism Trust remains the most frequently examined dimension in AI-journalism scholarship, appearing in 35 of the 46 reviewed studies. Two levels of trust consistently emerge: audience trust and journalistic trust. 4.3.1. Audience trust Most audience-facing studies reveal cautious acceptance of AI-generated content when transparency and editorial oversight are explicit. Wölker and Powell (2021) found that readers valued accuracy but still preferred stories authored or at least verified by humans. Similarly, Cools and Koliska (2024) observed that labeling AI-written articles improved trust only when accompanied by clear explanations of algorithmic logic. Across samples, disclosure without explanation was often counterproductive, heightening skepticism rather than reducing it (Karlsson and Clerwall, 2018). 4.3.2. Journalistic trust Inside the newsroom, trust refers to journalists’ willingness to rely on AI outputs in their daily routines. Jamil (2023) reported that while reporters appreciate AI’s speed and fact-checking support, they remain uneasy about its interpretive limits. Studies from the Reuters Institute (Beckett, 2019) and the Tow Center (Diakopoulos, 2019) show that AI adoption succeeds when staff perceive systems as augmenting—not replacing—editorial judgment. Taken together, these findings reinforce Trust Theory’s argument that confidence arises from perceived ability, benevolence, and integrity (Mayer et al., 1995). For AI in newsrooms, these qualities translate into technical reliability, ethical design, and institutional accountability. The review concludes that fostering both internal and external trust requires transparent communication about AI roles and human oversight mechanisms. 4.4. Theme 2: Editorial Accountability and Transparency Accountability represents journalism’s ethical backbone but is under increasing strain in algorithmic environments. Eighteen reviewed studies explicitly addressed accountability practices or transparency mechanisms in AI-driven journalism. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1068 4.4.1. Transparency as procedural accountability Karlsson and Clerwall (2018) conceptualize transparency as the audience’s window into journalistic processes. Many organizations have introduced “AI disclosure statements,” yet these remain inconsistent and vague. Vos and Craft (2017) argue that without explaining how algorithms influence content, such statements risk becoming symbolic rather than substantive. 4.4.2. Responsibility gaps The literature repeatedly highlights ambiguity over responsibility when errors occur. Beckett (2019) and Dörr (2016) noted that AI vendors, editors, and data scientists operate within different accountability logics, often lacking shared ethical standards. Plaisance (2015) and Jamil (2023) emphasize the need for distributed accountability—a framework recognizing shared moral agency among human and machine actors. 4.4.3. Governance implications Across institutional reports, a consensus emerges: governance frameworks must extend existing editorial codes to include algorithmic transparency, explainability, and bias auditing. UNESCO (2023) recommends mandatory impact assessments before deploying newsroom AI. This aligns with the IS Success Model’s service quality and net benefit dimensions, suggesting that ethical governance contributes directly to system success. 4.5. Theme 3: Workflow Transformation and Quality Workflow transformation studies (n = 28) focus on how AI alters newsroom organization, task distribution, and perceived quality of journalistic output. 4.5.1. Efficiency vs. editorial value AI tools demonstrably enhance efficiency automating routine updates, transcription, or data mining (Hansen et al., 2017; Dörr, 2016). Yet efficiency alone does not guarantee quality. Beckett (2019) observed that heavy reliance on automation may narrow editorial diversity, while Jamil (2023) found that reporters felt detached from creative storytelling when AI handled preliminary drafting. 4.5.2. Socio-technical balance STS theory predicts such tensions: introducing high-autonomy systems without redesigning social roles leads to mismatch and frustration (Mumford, 2006). Newsrooms that include journalists in AI-tool design—such as the BBC’s “Project Comma” initiative—report higher satisfaction and innovation (FIAT/IFTA, 2015). Thus, workflow quality depends on balancing technical affordances with human agency, echoing the socio-technical principle of joint optimization. 4.5.3. Skills and professional identity Multiple sources highlight the emergence of new hybrid roles “automation editors,” “AI curators,” and “data ethics officers.” These positions embody the organizational adaptation necessary for sustainable AI integration. Training and ethical literacy are increasingly recognized as success factors equal in importance to software performance (Cools and Koliska, 2024). 4.6. Cross-Theme Synthesis When analyzed together, the three themes reveal that newsroom AI’s success hinges on equilibrium among trust, accountability, and workflow quality. Overemphasizing one dimension undermines the others: technological efficiency without accountability erodes trust; excessive governance without usability hampers innovation. This interdependence justifies the 4-Dimensional (4D) Evaluation Framework proposed in Section 5, where each dimension Technical, Human-Organizational, Ethical-Governance, and Trust-Value is defined and operationalized. The framework responds directly to the deficiencies identified here and offers a structured foundation for future empirical validation. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1069 4.7. Proposed multi-dimensional evaluation framework The synthesis of theories and findings across 46 studies revealed that evaluating agentic AI in newsrooms requires a holistic, interdependent approach that balances technological performance with ethical, organizational, and social responsibility. Drawing from the Information Systems Success Model, Socio-Technical Systems Theory, Accountability Theory, and Trust Theory, this study proposes a 4-Dimensional (4D) Evaluation Framework that captures the full spectrum of AI influence in contemporary journalism. 4.8. Conceptual Overview Traditional evaluation frameworks measure AI success primarily through accuracy, efficiency, or user satisfaction. However, newsroom AI operates within a complex socio-technical and normative system where technology decisions intersect with human judgment and ethical standards. The proposed framework recognizes that success in this environment depends on four complementary dimensions: • Technical Quality – assessing functionality, reliability, and integration of AI systems. • Human-Organizational Alignment – capturing how journalists and editors interact with, adapt to, and cocreate with AI tools. • Ethical-Governance Responsibility – evaluating transparency, accountability, and compliance with professional norms. • Trust-Value Impact – measuring internal and external trust alongside the perceived public value of AI-assisted journalism. These four lenses together offer a structured, evidence-based approach to evaluating agentic AI performance in media environments. The framework’s conceptual model is illustrated in Figure 2. Figure 2 4D Evaluation Framework for Agentic AI in Newsrooms 4.9. Dimension 1 Technical Quality Technical quality remains foundational to any IS evaluation. In the context of agentic AI, it concerns the accuracy, speed, and stability of system performance as well as the transparency of algorithms. Drawing from DeLone and McLean’s (2016) “system quality” dimension, this level measures how well AI tools deliver on their intended editorial tasks such as data analysis, summarization, or story generation without introducing factual or contextual errors. Beyond traditional IT performance, newsroom AI quality also includes explainability the degree to which journalists can understand why the system produces certain outputs (Diakopoulos, 2019). Technical quality thus forms the foundation for trust, but it must be assessed alongside human and ethical considerations to prevent over-reliance on opaque automation. World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1076 Appendix Apendix 1 AI Journalism screening Matrix Author(s) Year Title / Source Journal / Institution Region / Focus Main Method Mapped Framework Dimension(s) Diakopoulos, N. 2017 Algorithmic Transparency in the News Media Digital Journalism Global Mixed Ethical–Governance; Trust– Value Dörr, K. N. 2016 Mapping the field of automated journalism Digital Journalism Global Conceptual Technical Quality; Human– Organizational Graefe, A. 2016 Guide to Automated Journalism Tow Center (Columbia University) Global Report Technical Quality; Human– Organizational Clerwall, C. 2014 Enter the Robot Journalist: Users' perceptions of automated content Journalism Practice Europe Experimental Trust–Value; Technical Quality Danzon-Chambaud, S. 2021 A systematic review of automated journalism scholarship Open Research Europe Global Systematic Review All Dimensions Siitonen, M. 2024 Mapping Automation in Journalism Studies 2010– 2019 Journalism Studies Global Review Human–Organizational; Ethical–Governance Wölker, A., & Powell, T. E. 2021 Algorithms in the Newsroom? Digital Journalism Europe Mixed Human–Organizational; Trust– Value Karlsson, M., & Clerwall, C. 2018 Transparency to the Rescue? Evaluating citizens' perceptions of transparency tools in journalism Journalism Studies Europe Mixed Ethical–Governance; Trust– Value Beckett, C. 2019 New powers, new responsibilities: A global survey of journalism and AI LSE Polis / JournalismAI Report Global Survey/Repo rt Human–Organizational; Ethical–Governance Diakopoulos, N. 2019 Automating the News: How algorithms are rewriting the media Harvard University Press Global Book Technical Quality; Trust–Value World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1077 Henestrosa, A. L. 2023 Automated Journalism: The effects of AI authorship and perceptions Journalism Studies Global Empirical Trust–Value; Human– Organizational Mooshammer, S. 2022 There are (almost) no robots in journalism Journalism Practice Global Critical Analysis Human–Organizational Norambuena, B. K. 2023 Using Transparency Cues to Help News Audiences Assess AI Media and Communication Global Experimental Ethical–Governance; Trust– Value Reuters Institute 2024 AI and the Future of News (research overview) Reuters Institute for the Study of Journalism Global Report All Dimensions Tow Center 2016 Guide to Automated Journalism Columbia University Global Report Technical Quality; Human– Organizational UNESCO 2023 Guidelines for AI in Journalism: Ethical and governance frameworks UNESCO Global Policy Report Ethical–Governance OECD 2021 OECD Principles on Artificial Intelligence OECD Global Policy Ethical–Governance; Trust– Value Plaisance, P. L. 2015 Media Ethics: Key principles for responsible practice SAGE Global Book Ethical–Governance DeLone, W. H., & McLean, E. R. 2016 Information systems success measurement Foundations and Trends in IS Global Conceptual Technical Quality Mumford, E. 2006 The story of socio-technical design Information Systems Journal Global Qualitative Human–Organizational Mayer, R. C., Davis, J. H., & Schoorman, F. D. 1995 An integrative model of organizational trust Academy of Management Review Global Conceptual Trust–Value Brennen, S., & Kreiss, D. 2016 Digitalization and journalism Digital Journalism Global Conceptual Technical Quality; Human– Organizational Dörr, K. N. 2019 Automated news in practice: a cross-national exploratory study Open Research Europe Global Empirical Human–Organizational Danzon-Chambaud, S. 2023 Automated news in practice: cross-national study Open Research Europe Global Empirical Human–Organizational World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1078 Trusting News (project) 2025 Audience experiments on AI disclosure and trust Trusting News / Report Global Experiment Trust–Value Diakopoulos, N., & Koliska, M. 2017 Algorithmic Transparency in News Production Digital Journalism Global Mixed Ethical–Governance Clerwall, C. 2015 Enter the robot journalist? The implementation of automated news Journalism Studies Europe Qualitative Technical Quality Gallego, C. 2020 Automated Journalism and Newsroom Practices Journalism Practice Global Case Study Human–Organizational Van Dalen, A. 2012 The algorithms behind the headlines Journalism Practice Global Conceptual Technical Quality Broussard, M. 2018 Artificial Unintelligence: How Computers Misunderstand the World MIT Press Global Book Ethical–Governance Lewis, S. C., & Westlund, O. 2015 Actors, Actants, Audiences, and Activities in the News Ecosystem Digital Journalism Global Conceptual Human–Organizational Fanta, A., & Dachwitz, I. 2020 Editorial automation in European newsrooms Otto Brenner Foundation Report Europe Report Human–Organizational Jamil, S. 2023 Artificial intelligence and journalism: Emerging trends, ethical dilemmas, and trust challenges Journalism Studies Africa/Glo bal Qualitative Human–Organizational; Ethical–Governance Moeller, S. 2020 Algorithmic curation and editorial control New Media & Society Global Qualitative Ethical–Governance; Human– Organizational LeCompte, K. 2021 AI-assisted investigative reporting: practices and ethics Investigative Journalism Review Global Case Studies Ethical–Governance; Technical Quality Kovach, B., & Rosenstiel, T. 2014 Elements of Journalism Crown Publishing Global Book Ethical–Governance Westlund, O. 2025 Digital Journalism (Studies): An Agenda for the Future Digital Journalism Global Conceptual Human–Organizational World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1079 Mooshammer, S. 2021 Automation taxonomy in journalism Journalism Studies Global Review Human–Organizational Siitonen, M. 2023 Automated journalism and institutional dynamics Journalism Studies Global Review Human–Organizational Seychell, D., et al. 2024 AI as a Tool for Fair Journalism: Case Studies from Malta arXiv / Preprint Malta Case Study Technical Quality; Ethical– Governance Yeung, W. N. 2024 Automated Journalism: Historical overview and critique arXiv / Preprint Global Review All Dimensions World Journal of Advanced Research and Reviews, 2025, 28(02), 1061-1080 1080 Apendix 2 PRISMA Flow Diagram