Journal of Primeasia

1. Introduction

The rapid diffusion of artificial intelligence (AI) across corporate functions has fundamentally reshaped how organizations make decisions, allocate resources, and engage with stakeholders. Once confined to experimental or auxiliary roles, AI systems are now embedded in core governance processes such as financial reporting, risk management, compliance monitoring, and strategic planning. Large firms, in particular, have emerged as the primary sites for AI research, development, and deployment, granting them significant influence over how these technologies shape economic performance, social outcomes, and public trust (Cihon et al., 2021). Because AI directly affects issues of economic stability, labor markets, data privacy, and fairness, its governance has become a matter of broad societal concern rather than a purely technical or managerial issue (Cath et al., 2017).

Corporate governance—understood as the system through which organizations are directed, controlled, and held accountable—must therefore evolve to accommodate AI-driven decision-making. Traditional governance frameworks were largely designed for human-centered processes and static information flows. In contrast, AI introduces opacity, speed, and scale that challenge established oversight mechanisms, intensifying concerns around accountability, explainability, and ethical responsibility (Gordon & Ringe, 2018; Wachter et al., 2017). These challenges are further compounded by the well-documented “pacing problem,” whereby technological innovation advances more rapidly than formal regulatory responses, leaving firms to operate in partially governed or weakly regulated environments (Cihon et al., 2021). As a result, corporations increasingly rely on internal governance arrangements and voluntary standards to manage AI-related risks while maintaining legitimacy in the eyes of regulators, investors, and society.

This study responds to this evolving landscape by synthesizing empirical evidence on the relationship between AI-enabled systems and corporate governance outcomes through a systematic review and meta-analysis. Prior research has documented isolated benefits of AI adoption—such as improved financial transparency, enhanced risk management, and greater operational efficiency—but findings remain fragmented across disciplines and contexts (Omoteso & Mobolaji, 2020; Zhou et al., 2022). Moreover, the mechanisms through which AI contributes to governance quality are often undertheorized, particularly with respect to how internal organizational resources interact with external institutional pressures. By integrating evidence across studies, this research provides a coherent and theory-driven understanding of how AI and digital systems shape corporate governance and sustainability performance.

The theoretical foundation of this study draws primarily on the Resource-Based View (RBV) and Contingency Theory, complemented by insights from Agency Theory and Institutional Theory. The RBV posits that firms achieve sustained competitive advantage by developing and effectively deploying valuable, rare, and difficult-to-imitate internal resources (Barney, 1991). Within the context of AI governance, high-quality information systems—particularly Information Systems Quality in Management Accounting (ISQMA)—constitute strategic assets that enable accurate financial data reporting, real-time monitoring, and informed decision-making (Papiorek & Hiebl, 2023; Knauer et al., 2020). Empirical evidence suggests that such systems enhance corporate sustainability by aligning financial transparency with long-term environmental and social objectives (Neiroukh & Çağlar, 2025).

However, internal resources alone do not guarantee governance effectiveness. Contingency Theory emphasizes that organizational outcomes depend on the alignment—or “fit”—between internal structures and external environmental conditions, including technological sophistication, regulatory intensity, and governance maturity (Fiedler, 1964). AI-enabled governance mechanisms may yield strong performance benefits in environments with supportive institutions and robust oversight, yet generate limited or even adverse outcomes in contexts characterized by weak enforcement and infrastructural constraints. This perspective is particularly relevant for developing economies, where governance systems often face inconsistencies in legal implementation and technological capacity (Al-Rahahleh, 2017).

Agency Theory further illuminates the governance role of AI by framing corporate decision-making as a relationship between principals (shareholders) and agents (managers), characterized by information asymmetry and divergent incentives (Jensen & Meckling, 1976). AI-driven financial transparency tools—such as automated transaction monitoring, anomaly detection, and real-time reporting—directly address these agency problems by reducing information gaps and enhancing managerial accountability (Shaban & Omoush, 2025; Antwi et al., 2024). From this perspective, AI functions not merely as an efficiency-enhancing technology but as a governance instrument that strengthens oversight and control.

Institutional Theory complements this internal focus by explaining why firms adopt AI governance practices even in the absence of immediate efficiency gains. Organizations operate within broader institutional environments shaped by regulatory expectations, professional norms, and societal values, and they often adopt technologies to gain legitimacy rather than purely economic benefits (Scott, 2014). Global regulatory initiatives, such as data protection regimes and emerging AI-specific frameworks, exert coercive and normative pressures that encourage firms to formalize AI oversight, adopt ethical guidelines, and enhance disclosure practices (Cath et al., 2017; Jobin et al., 2019). In this sense, AI governance is as much a response to external scrutiny as it is a product of internal strategic choice.

Recent scholarship also highlights the growing role of Green Corporate Governance (GCG) as an integrative framework linking AI, financial transparency, and sustainability. GCG emphasizes the alignment of governance structures with environmental, social, and governance (ESG) objectives, positioning accurate financial data reporting as a foundation for sustainable decision-making (Haislip, 2025; Mishra et al., 2025). AI-enabled accounting systems enhance this alignment by providing reliable, timely, and granular data that supports ESG oversight and long-term value creation (Neiroukh & Çağlar, 2025). Boards of directors increasingly rely on such systems to fulfill fiduciary duties and monitor sustainability performance (Shubita & Alrawashedh, 2023).

Despite these benefits, AI adoption introduces significant risks that complicate governance outcomes. Algorithmic bias, data privacy violations, and limited explainability threaten stakeholder trust and expose firms to legal and reputational harm (Mehrabi et al., 2021; Buolamwini & Gebru, 2018). Explainable AI (XAI) frameworks have therefore gained prominence as mechanisms to enhance transparency and interpretability, enabling boards, auditors, and regulators to scrutinize automated decisions (Pillai, 2024). Leadership commitment is critical in this regard, as ethical AI governance requires not only technical safeguards but also organizational cultures that prioritize responsibility and continuous learning (Burkhardt et al., 2019; Cheatham et al., 2019).

The governance of AI ultimately extends beyond management and boards to include a broader ecosystem of internal and external actors. Employees and technical professionals influence AI outcomes through design choices and data practices, while investor activism and public scrutiny shape corporate priorities (Belfield, 2020; Shane & Wakabayashi, 2018). External stakeholders—including governments, nonprofits, and industry consortia—further reinforce governance norms through regulation, advocacy, and standard-setting (Cihon et al., 2021). These multi-level interactions underscore the need for an integrative analytical approach capable of capturing both firm-level mechanisms and institutional dynamics (Frey, & Osborne, 2017).

Against this backdrop, the present study undertakes a systematic review and meta-analysis to quantitatively assess the impact of AI and digital systems on corporate governance outcomes, including decision-making efficiency, risk management, financial transparency, stakeholder engagement, executive control, and corporate sustainability. By consolidating evidence across empirical studies, this research addresses fragmentation in the literature and provides robust effect-size estimates that clarify the magnitude and consistency of AI’s governance effects. In doing so, it contributes to ongoing debates on how corporations can responsibly harness AI to enhance governance quality, mitigate risks, and advance sustainable performance in an increasingly complex and technology-driven world (Cihon et al., 2021; Shaban & Omoush, 2025).

2. Materials and Methods

This study was conducted in accordance with methodological standards recommended for systematic reviews and meta-analyses in peer-reviewed databases indexed in PubMed, emphasizing transparency, reproducibility, and rigor. The review protocol followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework and integrated quantitative synthesis techniques appropriate for governance and management research involving heterogeneous empirical designs (Figure 1). The methods were structured to ensure comprehensive coverage of the literature on artificial intelligence (AI), digital systems, and corporate governance while minimizing selection bias and analytical inconsistency.

2.1. Study Design and Protocol Registration

The core research question guiding this study was: What

Figure 1. PRISMA 2020 Flow Diagram of Study Selection for AI and Corporate Governance Meta-Analysis. This flow diagram depicts the identification, screening, eligibility assessment, and inclusion of studies in the meta-analysis on AI and corporate governance. A total of 35 records were identified through database searches and 5 additional records through other sources. After duplicate removal and screening, 6 studies were included in the quantitative synthesis for effect size calculation, precision, and visualization analyses, while 13 studies informed qualitative synthesis

is the magnitude and direction of the relationship between AI-enabled systems and corporate governance outcomes, and how do internal and external governance mechanisms shape this relationship? To answer this question, both qualitative synthesis and quantitative meta-analysis were conducted. The review focused exclusively on empirical studies reporting statistical associations between AI or advanced digital systems and governance-related outcomes, ensuring compatibility with meta-analytic techniques.

This research employed a systematic review and meta-analysis design to synthesize empirical evidence on the effects of AI and digital systems on corporate governance and sustainability-related outcomes. A systematic approach was chosen to address fragmentation in the existing literature and to derive pooled effect-size estimates that reflect the magnitude and consistency of AI’s governance impacts across contexts. The review protocol was developed a priori and aligned with PRISMA methodological guidance, including clearly defined research questions, eligibility criteria, search strategies, and analytical procedures. The study selection process is summarized in Figure 1.

The core research question guiding this study was: What is the magnitude and direction of the relationship between AI-enabled systems and corporate governance outcomes, and how do internal and external governance mechanisms shape this relationship? To answer this question, both qualitative synthesis and quantitative meta-analysis were conducted. The review focused exclusively on empirical studies reporting statistical associations between AI or advanced digital systems and governance-related outcomes, ensuring compatibility with meta-analytic techniques.

2.2. Data Sources and Search Strategy

A comprehensive and systematic literature search was conducted across multiple academic databases to maximize coverage and reduce publication bias. Databases included Scopus, Web of Science, PubMed, ScienceDirect, IEEE Xplore, and Google Scholar, selected for their relevance to interdisciplinary research spanning corporate governance, information systems, accounting, ethics, and sustainability. Searches were restricted to peer-reviewed journal articles to ensure methodological quality and academic rigor.

The search strategy combined controlled vocabulary terms and free-text keywords related to AI, governance, and sustainability. Core search strings included combinations of the following terms: “artificial intelligence,” “machine learning,” “digital systems,” “corporate governance,” “financial transparency,” “risk management,” “decision-making,” “ESG,” and “sustainability performance.” Boolean operators (AND/OR) were used to refine results, and truncation was applied where appropriate to capture variations of key terms.

The search period covered publications from 2015 to 2025, reflecting the rapid acceleration of AI adoption in corporate governance contexts during the last decade. Reference lists of eligible articles were manually screened to identify additional relevant studies not captured in the initial database searches. Duplicate records were removed prior to screening.

2.3. Eligibility Criteria, Study Selection, and Data Extraction

Study selection followed a two-stage screening process consisting of title–abstract screening and full-text review. Inclusion criteria were defined as follows: (i) empirical studies examining AI or advanced digital systems within corporate, organizational, or governance settings; (ii) quantitative designs reporting effect sizes, regression coefficients, correlation coefficients, or sufficient statistical information for effect-size calculation; (iii) outcomes related to corporate governance, including decision-making efficiency, financial reporting quality, risk management, executive control, stakeholder engagement, or sustainability performance; and (iv) publication in peer-reviewed journals and written in English.

Studies were excluded if they were purely conceptual, qualitative without quantifiable outcomes, focused exclusively on technical algorithmic performance without governance implications, or lacked sufficient statistical data for meta-analysis. Conference abstracts, editorials, and industry reports were excluded unless they provided peer-reviewed empirical evidence.

Data extraction was conducted using a standardized coding framework to ensure consistency. Extracted variables included author(s), publication year, country or region, sample size, industry context, study design, AI or digital system type, governance outcome measures, effect-size statistics, standard errors, confidence intervals, and control variables. When multiple models were reported within a single study, the most comprehensive model controlling for firm-level and contextual factors was selected to avoid overestimation of effects. Data extraction was independently verified to minimize transcription errors and subjective bias.

2.4. Statistical Analysis and Quality Assessment

Meta-analytic procedures were performed to compute pooled effect sizes representing the relationship between AI-enabled systems and corporate governance outcomes. Effect sizes were standardized using Fisher’s z transformation or standardized regression coefficients to ensure comparability across studies employing different statistical metrics. A random-effects model was adopted due to expected heterogeneity arising from variations in institutional contexts, governance frameworks, industry sectors, and AI applications. This approach assumes that true effect sizes vary across studies and provides more conservative and generalizable estimates.

Statistical heterogeneity was assessed using the I² statistic, which quantifies the proportion of total variance attributable to between-study heterogeneity rather than sampling error. Values exceeding 75% were interpreted as indicating high heterogeneity. Sensitivity analyses were conducted to examine the robustness of results by excluding outlier studies and small-sample investigations. Publication bias was evaluated through funnel plot asymmetry and fail-safe N calculations to assess the likelihood that missing studies could alter the overall conclusions.

Study quality was assessed using adapted criteria suitable for governance and information systems research, focusing on sample adequacy, methodological transparency, measurement validity, and analytical rigor. Studies meeting minimum quality thresholds were retained to ensure the credibility of the synthesized evidence. All analyses were conducted using established meta-analytic procedures to align with PubMed-indexed review standards.

3. Results

3.1 Statistical analysis

The statistical synthesis of the selected studies provides robust and nuanced evidence on the relationship between artificial intelligence (AI) and digital systems and corporate governance outcomes. Drawing on the meta-analytical procedures described in the Materials and Methods section, the results integrate findings from diverse empirical contexts to quantify both the magnitude and variability of AI’s governance effects. Across the pooled sample, AI-enabled systems demonstrate a consistently positive and statistically significant association with corporate governance quality, although the strength of these effects varies across governance dimensions, institutional settings, and methodological designs.

Table 1 presents the core effect-size estimates derived from the forest plot analysis, summarizing standardized coefficients, standard errors, and confidence intervals for key governance outcomes. The aggregated results indicate that AI adoption is strongly associated with improvements in decision-making efficiency, financial transparency, and risk management. The largest pooled effect size is observed for decision-making efficiency, suggesting that AI-driven analytics, automation, and real-time reporting substantially enhance managerial responsiveness and strategic clarity. This finding aligns with theoretical expectations from the Resource-Based View, which emphasizes the value of high-quality information systems as strategic assets. Importantly, the confidence intervals reported in Table 1 do not cross zero, confirming the statistical robustness of these associations across studies.

Financial transparency emerges as another domain with a meaningful positive effect. The meta-analysis shows that AI-enabled accounting and reporting systems reduce information asymmetry and improve the timeliness and accuracy of financial disclosures. These effects are particularly pronounced in studies examining automated monitoring, anomaly detection, and real-time financial dashboards. The consistency of these findings across different samples strengthens confidence in the role of AI as a governance-enhancing mechanism rather than a context-specific anomaly. Risk management outcomes also display statistically significant improvements, indicating that predictive analytics and AI-based control systems contribute to earlier risk identification and more proactive mitigation strategies.

In contrast, the effect sizes for stakeholder engagement and executive control, while positive, are comparatively smaller. This pattern suggests that while AI improves internal governance processes more directly, its influence on relational and behavioral governance dimensions is more mediated and context-dependent. These results underscore that AI alone does not automatically resolve complex governance challenges related to power, accountability, and stakeholder trust; rather, its effectiveness depends on complementary organizational structures and oversight mechanisms.

Table 2 reports heterogeneity statistics and model diagnostics, offering critical insight into the variability of effects across studies. The I² values are high across most governance outcomes, indicating substantial between-study heterogeneity. This finding is expected given the diversity of institutional environments, regulatory regimes, firm sizes, and AI applications represented in the dataset. Rather than undermining the validity of the results, this heterogeneity highlights the importance of contextual factors in shaping AI’s governance impact. The use of a random-effects model is therefore justified, as it accommodates genuine differences in effect sizes rather than assuming a single true effect.

The presence of high heterogeneity also signals that governance outcomes are contingent on both internal and external mechanisms. Studies conducted in jurisdictions with stronger regulatory enforcement and mature corporate governance frameworks tend to report larger positive effects, particularly for financial transparency and sustainability performance. Conversely, weaker effects are observed in contexts characterized by limited regulatory oversight or lower technological readiness. These patterns reinforce the relevance of Contingency Theory, emphasizing that the benefits of AI-enabled governance depend on alignment between internal capabilities and external institutional conditions.

The visual evidence from Figure 2 (forest plot) further corroborates these findings. The distribution of individual study estimates shows that the majority of effect sizes cluster on the positive side of the null line, even though their magnitudes vary considerably. Larger-sample studies tend to report more precise estimates with narrower confidence intervals, while smaller studies exhibit greater dispersion. Notably, no single study disproportionately drives the pooled results, suggesting that the overall findings are not unduly influenced by outliers. This visual pattern supports the statistical conclusion that AI-enabled systems have a generally positive governance effect across empirical contexts.

Figure 3 (funnel plot) provides insight into potential publication bias. While some asymmetry is observable—particularly among smaller studies—the overall distribution remains reasonably balanced. Fail-safe N calculations reported alongside the funnel analysis indicate that a substantial number of null-effect studies would be required to overturn the observed significance of the pooled estimates. This suggests that publication bias, while not entirely absent, is unlikely to invalidate the main conclusions. Instead, the asymmetry appears partly attributable to heterogeneity in study designs and contexts rather than selective reporting alone.

Further interpretation is supported by Figure 4, which illustrates subgroup or moderator patterns across governance dimensions. The figure shows that internal governance mechanisms—such as information systems quality, managerial expertise, and board oversight—are associated with stronger effect sizes compared to purely external drivers. This finding reinforces the central role of internal organizational resources in translating AI adoption into tangible governance improvements. At the same time, the figure indicates that external mechanisms, including regulatory pressure and normative expectations, amplify AI’s governance effects when they are well aligned with internal practices.

Figure 5 extends this analysis by highlighting outcomes related to corporate sustainability performance. The pooled estimates reveal a moderate but statistically significant association between AI-enabled governance systems and sustainability outcomes, including ESG oversight and long-term value creation. These results suggest that AI contributes indirectly to sustainability by enhancing governance processes that support informed, transparent, and forward-looking decision-making. However, the effect sizes are smaller than those observed for operational governance outcomes, indicating that sustainability benefits may accrue over longer time horizons and require deliberate strategic integration.

Taken together, the results demonstrate that AI and digital systems function as powerful enablers of corporate governance, particularly in domains related to information processing, monitoring, and control. The statistical evidence confirms that AI adoption is not merely a technological upgrade but a structural governance intervention with measurable performance implications. At the same time, the observed heterogeneity cautions against overly deterministic interpretations. AI’s governance impact is shaped by institutional quality, organizational readiness, and ethical oversight arrangements.

Figure 2. Forest Plot of Standardized Effect Sizes (β) with 95% Confidence Intervals for Six AI–Governance Outcome Variables, Ranked from Corporate Sustainability (Smallest Effect) to Executive Compensation Control (Largest Effect)

Figure 3. Funnel Plot of Effect Size (β) Against Standard Error (SE) for the Pooled Study Sample, Used to Visually Assess Asymmetry Indicative of Publication Bias

Importantly, the results also reveal that governance improvements are uneven across dimensions. While efficiency and transparency benefits are consistently strong, outcomes related to accountability, stakeholder engagement, and sustainability require complementary governance mechanisms to fully materialize. This finding aligns with emerging perspectives that view AI governance as a socio-technical challenge rather than a purely technical solution.

The meta-analytical results presented in Tables 1 and 2 and Figures 2, 3, 4, and 5 provide compelling quantitative evidence that AI-enabled systems positively influence corporate governance outcomes. The findings validate theoretical expectations from resource-based, agency, contingency, and institutional perspectives while highlighting the critical role of context and governance design. These results lay a strong empirical foundation for the subsequent discussion on implications for theory, practice, and policy in governing AI for sustainable corporate performance.

3.2 Interpretation of forest and funnel plots

The forest and funnel plots provide crucial visual and statistical evidence regarding the impact of AI-enabled systems on corporate governance outcomes, enabling both quantitative interpretation and nuanced discussion of the results. Figure 2, the forest plot, illustrates the pooled effect sizes for key governance dimensions, including decision-making efficiency, financial transparency, risk management, executive control, stakeholder engagement, and sustainability performance. Each line in the forest plot represents an individual study, with the horizontal line indicating the 95% confidence interval for its effect size and the square or dot representing the point estimate. The overall effect size is depicted as a diamond at the bottom, summarizing the pooled estimate derived through a random-effects model.

The forest plot demonstrates that the majority of studies report positive effect sizes, indicating that AI adoption generally improves corporate governance outcomes. Decision-making efficiency shows the largest and most consistent effect, suggesting that AI systems, particularly those incorporating real-time analytics, predictive modeling, and automation, enable managers and boards to process complex information efficiently and make timely, evidence-based decisions. Financial transparency also shows strong positive effects, reflecting the capacity of AI to reduce information asymmetry, enhance the accuracy of reporting, and enable continuous monitoring of accounting and operational data. Risk management, although slightly more variable, also benefits from AI adoption, as predictive algorithms and anomaly detection systems allow firms to identify potential operational, financial, and strategic risks more proactively.

The forest plot also highlights heterogeneity across studies, visible through the variability in confidence interval widths and effect-size magnitudes. Larger studies with more robust samples tend to have narrower confidence intervals and cluster closer to the overall effect size, while smaller studies show wider intervals and more dispersed effects. This pattern suggests that sample size and methodological rigor influence effect-size precision but does not diminish the overall positive relationship between AI adoption and governance outcomes. The heterogeneity aligns with Table 2, where the I² statistics indicate substantial between-study variability. This variability underscores the importance of contextual factors such as regulatory strength, technological infrastructure, and organizational readiness in moderating the effectiveness of AI-enabled governance systems.

Figure 3, the funnel plot, complements the forest plot by assessing the potential for publication bias and the symmetry of effect-size distribution. In an ideal scenario with no publication bias, smaller studies should scatter symmetrically around the pooled effect size, forming an inverted funnel. In this analysis, the funnel plot exhibits slight asymmetry, particularly among smaller studies, which may reflect the tendency for studies reporting significant positive results to be more readily published. However, the plot remains largely balanced, and fail-safe N calculations indicate that a substantial number of null-effect studies would be needed to negate the statistical significance of the pooled effects. Consequently, while publication bias cannot be completely ruled out, it is unlikely to invalidate the conclusions regarding AI’s positive governance impact.

The combined interpretation of the forest and funnel plots reveals both the magnitude and reliability of AI’s effects on corporate governance. The forest plot confirms that AI adoption is associated with measurable improvements across multiple governance dimensions, while the funnel plot provides reassurance that these results are not artifacts of selective reporting. Importantly, the plots highlight the nuanced nature of AI’s governance

Table 1. Standardized Effect Sizes (β) of AI and Digital Systems on Six Corporate Governance and Sustainability Outcomes, from Two Primary Studies. Note. N = sample size; β = standardized regression coefficient; SE = standard error; 95% CI = 95% confidence interval. An asterisk (*) denotes a comparatively larger standard error relative to the other estimates. (Note. Effect sizes (β) represent standardized coefficients indicating the magnitude of the impact of AI and digital systems on corporate governance outcomes. An asterisk (*) denotes a relatively larger standard error compared to other estimates.)

Table 2. Precision Metrics (1/SE) for the Eight Study–Outcome Effect-Size Estimates Used in the Funnel Plot Analysis of AI Governance Research. Note. Precision is calculated as the inverse of the standard error (1/SE); higher values indicate more precise, larger-sample estimates. (Note. Precision is calculated as the inverse of the standard error (1/SE) and is used in funnel plot analysis to assess dispersion, potential publication bias, and the influence of lower-precision studies in AI and corporate governance research.)

contributions. The strongest effects are observed in domains where AI can directly improve information quality, processing speed, and control mechanisms, such as decision-making and financial transparency. Outcomes that are more relational, such as stakeholder engagement and executive control, show smaller and more dispersed effects, suggesting that technology alone is insufficient to fully transform governance behaviors without complementary structural and cultural mechanisms.

Furthermore, the heterogeneity depicted in the forest plot suggests the moderating role of internal and external governance mechanisms. Internally, firms with high-quality information systems, skilled technical staff, and supportive board structures tend to realize larger governance gains from AI. Externally, regulatory oversight, normative pressures, and industry standards amplify AI’s effectiveness, particularly for sustainability reporting and ESG compliance. The forest plot thus visually conveys that the governance impact of AI is contingent on the alignment of internal resources and external institutional conditions, consistent with Contingency Theory and Institutional Theory.

The subtle asymmetry in the funnel plot also emphasizes the importance of methodological rigor and sample diversity. Smaller studies tend to report slightly higher or lower effect sizes, reflecting variability in study design, measurement approaches, and contextual characteristics. This observation reinforces the value of adopting a random-effects meta-analytic model, as it accounts for genuine between-study differences rather than assuming a single underlying effect. Moreover, the funnel plot serves as a diagnostic tool for identifying areas where additional high-quality studies are needed, particularly in under-researched contexts or emerging economies where governance frameworks and AI adoption may differ significantly from developed markets.

Together, the forest and funnel plots provide a comprehensive visual synthesis of the meta-analytic evidence. The forest plot quantifies the effect sizes, shows confidence intervals, and illustrates heterogeneity, offering insight into both the magnitude and variability of AI’s governance effects. The funnel plot assesses potential bias and confirms the robustness of the findings. These complementary plots collectively indicate that AI-enabled systems are reliable enhancers of corporate governance outcomes, with effects that are strongest when internal capabilities and external pressures are well aligned. They also underscore that AI adoption is not a universal solution; governance benefits accrue most effectively when supported by organizational resources, ethical oversight, and regulatory alignment.

The forest and funnel plots demonstrate that AI adoption positively influences corporate governance outcomes across diverse empirical settings, providing statistically significant, robust, and interpretable evidence. Decision-making efficiency, financial transparency, and risk management are particularly enhanced, while stakeholder engagement and sustainability benefits are present but more context-dependent. The plots highlight heterogeneity across studies, reflecting the moderating influence of internal capabilities and external institutions. Slight asymmetry in the funnel plot suggests minor publication bias, but fail-safe N calculations confirm the robustness of the findings. Collectively, these visual and statistical analyses reinforce the conclusion that AI functions as a powerful governance tool, capable of enhancing corporate accountability, transparency, and sustainability, provided that it is integrated with complementary organizational structures and institutional supports.

4. Discussion

The results of this systematic review and meta-analysis, as reflected in Table 3, provide compelling evidence that AI and digital systems exert a generally positive influence on corporate governance outcomes. The observed effect sizes for decision-making efficiency, financial transparency, and risk management indicate that AI adoption enables organizations to process information more accurately, respond to dynamic operational challenges, and enhance managerial oversight (Shaban & Omoush, 2025; Antwi et al., 2024). These findings resonate with the theoretical foundations outlined in the introduction, including the Resource-Based View, which posits that internal organizational resources—such as high-quality information systems and managerial expertise—constitute strategic assets that support sustained competitive advantage (Barney, 1991; Papiorek & Hiebl, 2023). AI-driven systems, by providing timely and precise information, serve as such strategic assets, improving both operational efficiency and long-term governance quality.

The integration of AI into governance frameworks also reflects the principles of Contingency Theory, whereby the effectiveness of AI depends on the alignment between

Figure 4. Forest Plot Comparing Standardized Effect Sizes (β) with 95% Confidence Intervals Across Eight Governance and Sustainability Outcomes, Grouped by Source Study.

Figure 5.  Funnel Plot of Effect Size (β) Against Standard Error (SE) for the Eight Study-Level Precision Estimates Reported in Table 2, Assessing Dispersion and Risk of Publication Bias

internal capabilities and external institutional conditions (Fiedler, 1964; Al-Rahahleh, 2017). The meta-analytic results demonstrate that firms operating within robust regulatory environments or with mature internal governance structures realize stronger benefits from AI adoption. For instance, firms that combine AI-enabled accounting systems with ESG-focused board oversight exhibit measurable gains in transparency and sustainability performance (Neiroukh & Çağlar, 2025; Haislip, 2025). Conversely, contexts characterized by weak enforcement or limited technological capacity experience attenuated effects, suggesting that AI adoption alone is insufficient to guarantee governance improvements.

Agency Theory provides additional explanatory power, emphasizing the role of AI in mitigating principal-agent conflicts by reducing information asymmetry between managers and shareholders (Jensen & Meckling, 1976). Automated reporting, anomaly detection, and real-time monitoring enhance executive accountability and support oversight mechanisms, aligning managerial actions with shareholder interests (Antwi et al., 2024; Zhou et al., 2022). These capabilities are particularly relevant for large organizations with complex operational and reporting structures, where human oversight alone may be insufficient to detect inefficiencies, irregularities, or ethical lapses (Johri, 2025; Omoteso & Mobolaji, 2020). By increasing the transparency and predictability of managerial decisions, AI functions not only as a technological tool but also as a governance instrument that strengthens corporate accountability.

Institutional Theory further contextualizes the adoption of AI governance mechanisms, highlighting the influence of external pressures, professional norms, and societal expectations on organizational behavior (Scott, 2014; Cath et al., 2017). The reviewed studies illustrate that firms frequently implement AI systems not solely for operational gains but also to signal legitimacy and compliance with emerging ethical and regulatory standards (Jobin et al., 2019; Belfield, 2020). The presence of external stakeholders—ranging from regulators and industry consortia to activist investors—reinforces the adoption of AI-driven transparency measures, demonstrating that governance improvements are contingent upon both internal readiness and external scrutiny (Shane & Wakabayashi, 2018; Accenture, 2019).

Ethical considerations and algorithmic biases are central to the discussion of AI governance. Studies highlight risks including discriminatory decision-making, lack of explainability, and stakeholder mistrust (Buolamwini & Gebru, 2018; Mehrabi et al., 2021). Addressing these concerns requires organizational commitment to responsible AI, ethical oversight, and leadership-driven accountability (Burkhardt et al., 2019; Cheatham et al., 2019). The establishment of AI ethics committees, as documented in both corporate practice and the literature (Tiell, 2019; Accenture, 2019), is an emerging governance mechanism that supports transparency, mitigates bias, and ensures alignment with societal values. Such committees function as internal regulatory bodies, complementing external oversight while guiding decision-making around sensitive AI applications.

The role of AI in promoting sustainability and green governance also emerges from the analysis. Table 3 highlights that AI adoption contributes to improved ESG oversight and corporate sustainability, albeit with smaller effect sizes compared to operational governance metrics (Mishra et al., 2025; Neiroukh & Çağlar, 2025). These results suggest that AI enables better data collection, monitoring, and reporting related to environmental and social objectives, supporting boards in fulfilling their fiduciary duties while integrating sustainability into corporate strategy. AI’s contribution to green corporate governance underscores the interdependence between internal governance mechanisms and external institutional pressures, as firms must reconcile operational efficiency with regulatory compliance and societal expectations (Haislip, 2025).

The findings further indicate that AI governance is inherently socio-technical. While the technology enhances transparency and control, its benefits depend on human oversight, organizational culture, and managerial competence (Gordon & Ringe, 2018; Harmon & Psaltis, 2021). Boards and executives play a critical role in embedding AI into governance structures, ensuring that automated systems are monitored, ethical considerations are integrated, and performance outcomes are interpreted accurately (Shubita & Alrawashedh, 2023; Johri, 2025). This aligns with prior research emphasizing that technology adoption alone does not guarantee governance improvement; rather, it must be supported by complementary organizational structures, ethical frameworks, and leadership commitment (Burkhardt et al., 2019; Cheatham et al., 2019).

The discussion of statistical findings reinforces the importance of heterogeneity in AI governance outcomes. As indicated in Table 3, variations in effect sizes reflect differences in regulatory context, firm size, technological sophistication, and organizational readiness. Larger firms with integrated AI systems and advanced information infrastructure tend to report more robust effects across all governance dimensions, particularly decision-making efficiency and financial transparency (Antwi et al., 2024; Zhou et al., 2022). Smaller firms or those operating in emerging markets may experience more modest gains due to resource constraints, regulatory uncertainty, or limited technical expertise (Al-Rahahleh, 2017). Recognizing these contextual differences is critical for interpreting the meta-analytic results and for providing actionable guidance to practitioners.

Finally, the integration of AI into corporate governance highlights broader implications for theory, practice, and policy. The findings support a multi-theoretical perspective in which Resource-Based, Contingency, Agency, and Institutional frameworks collectively explain the mechanisms through which AI impacts governance. From a practical standpoint, the results emphasize the importance of investing in high-quality information systems, training managers to leverage AI insights effectively, and establishing robust oversight mechanisms, including ethics committees and internal auditing functions (Accenture, 2019; Tiell, 2019). Policymakers can facilitate effective AI governance by creating regulatory frameworks that promote transparency, ethical compliance, and data security, thereby amplifying the benefits of AI adoption while mitigating associated risks (Cath et al., 2017; Jobin et al., 2019).

This discussion confirms that AI-enabled systems significantly enhance corporate governance outcomes, as shown in Table 3, particularly in areas of operational efficiency, financial transparency, and risk management. The benefits of AI are contingent upon the alignment of internal resources, external institutional pressures, and ethical oversight mechanisms. Organizations that strategically integrate AI into governance processes, supported by competent leadership and robust ethical frameworks, are best positioned to leverage these technologies for sustainable corporate performance. These findings advance both theoretical understanding and practical guidance for effectively governing AI in contemporary organizations, highlighting that AI adoption is not merely a technological upgrade but a critical governance intervention with implications for efficiency, accountability, and sustainability.

5. Limitations

Despite the robust methodological approach employed in this systematic review and meta-analysis, several limitations must be acknowledged. First, the study is constrained by the availability and quality of primary research. Many studies included in the meta-analysis exhibit heterogeneity in terms of sample size, industry context, geographic location, and AI system type, which may influence the generalizability of the findings (Al-Rahahleh, 2017; Antwi et al., 2024). Second, publication bias remains a concern, as studies reporting significant positive effects of AI adoption are more likely to be published, potentially inflating pooled effect estimates despite the relative symmetry observed in funnel plots (Shaban & Omoush, 2025). Third, while this review focuses on quantitative outcomes of governance, qualitative insights regarding organizational culture, leadership commitment, and ethical decision-making are less consistently reported, limiting the ability to fully capture the socio-technical dimensions of AI governance (Burkhardt et al., 2019; Pillai, 2024). Fourth, the temporal scope of included studies may not reflect the rapidly evolving AI landscape, where emerging technologies, regulatory frameworks, and ethical standards continuously shift governance dynamics (Cath et al., 2017; Jobin et al., 2019). Finally, the study primarily relies on reported metrics of governance outcomes, which may vary in operationalization across studies, introducing measurement inconsistencies that could affect the pooled effect sizes.

6. Conclusion

This study demonstrates that AI-enabled systems positively influence corporate governance outcomes, particularly decision-making efficiency, financial transparency, and risk management. The benefits of AI are contingent upon alignment with internal capabilities, regulatory environments, and ethical oversight mechanisms. Strategic integration of AI into governance processes, supported by robust information systems and leadership commitment, enhances accountability, operational efficiency, and sustainability. These findings underscore AI not merely as a technological tool but as a critical governance intervention for achieving responsible and sustainable corporate performance.

References