Journal of International DBA Studies - GGU

Governing AI Agents: Bounded Autonomy and Human Oversight

Prasad Modali — Thu, 30 Apr 2026 00:00:00 +0000

The sudden arrival of Artificial Intelligence (AI), is shifting both scholarly and governmental focus away from issues of technological feasibility to greater concern regarding its use and oversight. AI, which promises great advancements in operational efficiency, scalability and adaptive learning, however presents many challenges as it is used in large-scale applications, including those of ethics and compliance with law; accountability and the ability of humans to exercise meaningful control. Traditional decision-support systems differ from AI agents because they are capable of independently initiating action, coordinating workflow processes, and modifying behavior based upon past experience. These characteristics create new obstacles for existing governance and oversight mechanisms. Research continues to emphasize technical performance and model-based protections against unwanted behavior by AI agents. There remain substantial gaps in governance structures that can support autonomous AI while maintaining some level of institutional accountability. This paper proposes a Bounded Autonomy Governance Framework for AI Agents to be developed using autonomy as a design variable to be governed during all phases of the development process. The proposed framework includes several key components such as calibrated levels of autonomy for AI agents, human-in-the-loop and human-on-the-loop mechanisms to include ethical guardrails, and compliance-by-design provisions throughout the entire life cycle of an AI agent. Using a multi-disciplinary body of literature relating to AI governance, digital infrastructure, and organizational systems, the author views AI agents as social and technical actors embedded in institutions. The paper also builds from the bounded rationality and absorptive capacity literature to strengthen the theoretical case for studying bounded autonomy. Through both theoretical and practical design considerations —including an applied case walkthrough—this paper provides actionable recommendations for policymakers, platform designers, and industry practitioners wishing to responsibly develop and implement AI agents at scale.

AI-Enabled Training Micro-Agents Longitudinal Effects on Adoption, Learning Efficiency, and Human Oversight

Smrite Goudhaman — Wed, 25 Mar 2026 00:00:00 +0000

This study reports a longitudinal assessment of micro-agents facilitated by AI in a front-line hospitality environment. It compares a supervised micro-agent deployment in 2024 with a scaled organizational deployment in 2025. Deployment maturity is the independent variable, while objective learning platform trace metrics - completion rate, assessment performance, and time-on-task - that comprise the dependent variables of this study. An exposure-adjusted active employee model is utilized to reduce potential biases due to a high employee turnover rate. A decrease in completion rate is found from 100% (656/656) in the supervised micro-agent deployment to 86.82% (8,505/9,796) in the scaled micro-agent deployment. A two-proportion z-test shows a significant difference (z = 11.52, p < .001) with a decrease in completion rate by 13.18% (95% CI [12.51, 13.85]). This decline is consistent with normalization effects commonly observed when controlled pilot interventions transition to scaled operational environments. Assessment performance increases from M = 80.10 (SD = 21.55) to M = 83.38 (SD = 23.14) with a small effect size (Welch's t (1108.70) = 4.29, p < .001; Cohen's d = 0.14; 95% CI [1.78, 4.79]). A decrease in mean time-on-task is found from 10.30 minutes (SD = 11.53) to 5.98 minutes (SD = 7.82) with a moderate efficiency effect (Welch's t(971.64) = -10.88, p < .001; Cohen's d = -0.52; 95% CI [-5.09, -3.53]). The findings provide empirical evidence regarding the effectiveness of AI-enabled micro-agent frameworks within frontline organizational learning environments. It links longitudinal behavioral traces with micro-agent frameworks, providing a replicable model for assessing the effectiveness of AI-enhanced organizational learning systems.

Reciprocal Enablement of Data Centers and AI Agents: From Silicon Foundations to Sentient Operations

Ratheesh Venugopal — Wed, 25 Mar 2026 00:00:00 +0000

Artificial intelligence (AI) agents and data centers are in a symbiotic relationship of mutual enablement. AI agents and autonomous goal-oriented systems, which are able to perceive, reason, and act, are becoming more and more coordinated and specifically as related to mechanical, electrical, controls, and IT (MECIT). Simultaneously, hyperscale and edge data centers deliver the silicon, networks, storage hierarchies, thermal envelopes, and governance needed to scaffold and enable agentic systems to operate at scale in real time. To address this newer phenomenon, this article consolidates a cross-knowledge base (computer science, operations research, energy systems, and international business policy) and cross checks it with current industry facts to (a) explain the processes through which AI agents achieve efficiency, resiliency, sustainability, and security of data centers; (b) analyze how data center structures, supply chains, and instituting frameworks facilitate increasingly capable AI agents; and (c) appraise managerial, financial, and policy implications over the global digital infrastructure. Examples of case studies related to reinforcement-learning (RL) cooling optimization, carbon-aware scheduling, liquid-cooled AI clusters, multi-agent enterprise orchestration, and carrier-neutral interconnection fabrics are provided. Through these sample cases, we posit that agentic automation and carbon-conscious compute are complementary and facility-scale innovations (liquid cooling, power-dense racks, and edge-to-cloud fabrics). The paper ends with research and practice agenda implications based on quantifiable KPIs (e.g., PUE, WUE, partial PUE, embodied carbon per server, outage rates) and governance anchors (NIST AI RMF, ISO/IEC 42001, EU AI act).

A Dual Strategy for Digital Market Integrity: Content Credentials and Consumer Trust

Janak Makwana — Tue, 07 Apr 2026 00:00:00 +0000

The rapid advancement of generative AI has raised concerns about visual trust in digital commerce, as highly realistic synthetic images make it challenging to distinguish between authentic and artificial content. This article examines implications for e-commerce and online food delivery, using India’s fast-growing digital marketplace as context and extending to global policy trends. Grounded in signaling theory, it argues that AI-generated visuals act as low-cost, deceptive signals in information-asymmetric environments, eroding consumer confidence and inflating post-purchase dissatisfaction. Detection-only strategies fall short due to adversarial adaptation and model drift. It proposes the Verifiable Authenticity Framework: a phased playbook that (1) embeds Content Credentials via the C2PA standard across the media supply chain; and (2) ensures consumer-facing transparency at purchase, with optional cryptographic notarization for high-value categories. This enables brands to send costly, auditable signals of authenticity, restoring trust in image-driven markets. An experimental design is outlined to assess its impact on return rates, ratings, and repurchase intent. By embedding provenance and transparency, firms can reduce operational risk, align with emerging cross-jurisdictional mandates, and build sustainable competitive advantage in the global AI-driven economy.

Multimodal Generative AI Agents for Biomedical Document Classification: Architecture, Ethical Boundaries, and Human-in-the-Loop Governance

Arun Kumar — Mon, 20 Apr 2026 00:00:00 +0000

The rapid increase in biomedical research publications has made it difficult for researchers, clinicians, and policymakers to efficiently review and interpret scientific information. Traditional manual review methods and rule-based automation tools are no longer sufficient to manage the growing volume, complexity, and multimodal nature of modern biomedical literature, where important insights need to be presented through both written text and visual elements such as figures and images. To address this challenge, this study proposes and evaluates a multimodal generative AI agent for biomedical document classification and image captioning, which combines an instruction-tuned language model with a vision encoder to process abstract text and related visual content together. The agent operates within a controlled framework that includes human oversight to ensure responsible and ethical use. To study its effect a mixed-method approach was used, including quantitative performance evaluation and qualitative expert review. The model was tested on open-access biomedical papers from arXiv across four subject areas. Results indicate that the multimodal approach performs better than text-only systems in classification accuracy and contextual understanding. However, the findings also show that human supervision remains important in order to reduce risks related to bias and incorrect outputs. This study therefore offers practical and theoretical guidance for developing ethical and reliable AI systems in biomedical research settings.

From AI Hype to Agentic Reality: A Readiness Lens for Sustainable Enterprise Adoption

Richa Srivastava — Sat, 18 Apr 2026 00:00:00 +0000

This conceptual article examines organizational readiness as a critical but under-theorized condition for sustainable enterprise adoption of artificial intelligence (AI) agents. While AI agents are increasingly touted as the next stage of enterprise AI, many organizations are pursuing deployment without adequately preparing the governance, accountability, and organizational conditions that are required for responsible scale. This conceptual paper investigates the above by relying upon socio-technical systems theory, dynamic capabilities, and accountability theories needed to explore the organizational preparedness as a critical, but also as an under-theorized state of enterprise adoption of AI agents. As such, this paper introduces a new term: organizational readiness debt and defines it as an accrual of hidden governance, accountability, and legitimacy risks incurred due to the hasty implementation of agentic systems. To address this gap, the paper also includes the introduction of the Agentic Organizational Readiness Framework (AORF). AORF is a multi-dimensional diagnostic framework, which consists of strategic, governance, risk, workforce, architectural, and ethical-legitimacy dimensions. Using an integrative conceptual review and illustrative case-based analytic examples, the paper argues that organizational readiness, rather than technical capability alone, is a key determinant of agentic outcomes. As such, the paper contributes to the emerging literature on AI agents and offers practical guidance for senior leaders responsible for enterprise AI transformation.

Digital Public Infrastructure and Agentic AI: Shaping the Future of Insurance in India

Siddhartha Pappala — Mon, 27 Apr 2026 00:00:00 +0000

The Indian insurance sector is in the process of structural transformation of traditional agent-based distribution to technology-based ecosystem (Atlantic Council, 2025; Deloitte, 2025; EY, 2025; KPMG, 2025; Sharma et al., 2024). This conceptual paper focuses on the key contribution to this change represented by Agentic Artificial Intelligence (Agentic AI) systems in which decisions are made autonomously, adapt continuously, and are goal-oriented (He and Zhu, 2025; Hosseini et al., 2025; Sapkota et al., 2025). Rather than presenting empirical findings, the paper synthesizes secondary sources such as peer-reviewed literature, regulatory reports, and industry reports to develop a conceptual analysis of the transformation of four fundamental insurance value chain functions by Agentic AI: sales, underwriting, claims management, and policy servicing The analysis is designed by using Sense-Insight-Dynamics (SID) framework, a new framework that builds on the classical scheme of Perception-Cognition-Action agent framework (Wooldridge and Jennings, 1995; Russell and Norvig, 2016) by integrating in domain-specific insurance contexts. The paper explores how India Digital Public Infrastructure (DPI) has a fundamental role in facilitating scalability of these smart systems (Atlantic Council, 2025; Harvard Kennedy School, 2025; ORF America, 2025). There is also a critical evaluation of implementation issues such as algorithmic bias in the socio-culturally stratified society of India, integration with the old system, the overstaffed local population, lack of digital literacy in rural regions, and consent fatigue in DPI frameworks. The results are added to the increasing literature on AI-based financial inclusion in developing markets.

The Rise of AI and AI Agents: Evolution, Disruption, and Governance

Abigail Modino, Joaquin Gonzalez, Nicole Jackson — Mon, 11 May 2026 00:00:00 +0000