Jul 2026· Annual International Computer Software and Applications Conference· pp. 2575-2580· 0 citations· 16 references
Abstract
Most vulnerability pipelines remain predictioncentric: they output scores or labels and defer decisions to engineers, even when outputs are compressed, imbalanced, or unreliable under Continuous Integration and Continuous Deployment (CI/CD) shift. We introduce AEGIS (Autonomous Enhanced Guardian for Intelligent Security), a policy-governed framework that treats vulnerability management as a constrained CI/CD decision process in which learned signals serve as evidence and are translated into admissible actions under explicit constraints. AEGIS combines graph-based risk estimation, epistemic uncertainty via stochastic inference, and a symbolic policy guard that maps evidence to auditable decisions: Block, Warn, and Pass. A key design principle is separation of concerns: perception estimates risk while governance determines admissible actions. Irreversible automation is permitted only when risk is high and uncertainty is low, while uncertain cases are deferred to controlled review. This separation makes it possible to revise policy thresholds, cost assumptions, and review budgets without retraining the perception model. We evaluate AEGIS on an extreme-imbalance patch stream used as a stress-test setting and an expanded multi-project dataset that enables more stable estimation of decision outcomes. The evaluation reports the policy thresholds, model settings, symbolic predicates, ablations, and sensitivity settings used in policy replay. Results provide preliminary evidence that policygoverned control supports more interpretable decision behavior under uncertainty, while enabling controlled trade-offs between automation, safety, and review load.
The findings show that LLMs can approximate structured cybersecurity reasoning under controlled representations, but do not apply it robustly, which has important implications for the design and evaluation of AI-assisted security decision-support systems.
With the increasingly aggressive cyber threat landscape for governments, businesses, and institutions, as information and/or cybersecurity implementations are increasingly under scrutiny by regulators, it has been pointed out that governance failure is one of the major reasons for a weakened cybersecurity posture. A major component of Cyber/information security governance is the development, adoption, and implementation of a comprehensive information and/or cyber security policy document. The policy document must be in compliance with international or national standards and, if possible, with regulatory guidelines. However, it is often observed that policy documents are often incomplete with respect to industry standards or regulations and require revision when subjected to a thorough audit. Identifying the gaps between the controls and processes documented in the policy and those required in the regulations or standards necessitates extensive manual effort. The advent of Generative AI tools such as Large Language Models (LLMs) led to use of LLMs and Agentic AI tools to automate such compliance checks, as seen in a few research publications in recent times. However, such reported use of LLMs are experimented with high resource environments such as expensive GPUs and memory based servers. For smaller organizations such expensive compute platform may not be easily available. In this article, we benchmark the compliance checking tasks on LLMs that do not require GPU and high memory usage and the effectiveness of such resource constrained LLMs in compliance checking. Our experiments demonstrated that the low resource LLMs can provide good agreement/accuracy in compliance checking of policy documents against standards by experimenting with ISO 27002:2022 controls against multiple policy documents.
R. Negi, Rishika Jain, Soumyo V Chakarborty et al.· 0 citations
The high implementation of remote working models has only increased the pace of using artificial intelligence tools by employees to improve their productivity, automate routine tasks, and aid in decision-making. Non-committal application of AI tools, or Shadow AI, has, however, become a major organizational risk, especially when the processed information is sensitive, proprietary, or controlled and not handled in line with the existing organizational policies. The current mitigation is based on fairly flat policies, manual checks, or limited monitoring systems, which cannot provide real-time awareness, dynamic risk response, or gated controls within the dynamic remote workforce. To minimize the leaking of proprietary data, this study suggests a single framework of Shadow AI risk mitigation to detect, evaluate, and control unsanctioned AI utilisation. The architecture incorporates automated scoring of risks, policy-sensitive enforcement, risk uncertainty estimation, and governance-driven adaptation as a way of balancing security, usability, and operational efficiency. Experimental results based on synthetic enterprise data (20,000 activity records) and real-world organizational analytics (7,500 interaction records) show that it can detect jobs with an accuracy of 91.6, reduce high-risk Shadow AI interactions by a factor of 62.6 and reduce policy violations by half, using less than 16% computation overhead. The suggested solution offers a scaling- and governance-based solution to securing the use of AI solutions in distributed working scenarios.
A. Assiri· International Journal of Adv...· 0 citations
The deployment of artificial intelligence systems in enterprise risk management presents a novel governance challenge: determining which AI-generated risk assessments carry sufficient confidence for automated action and which warrant human review. This is the confidence-conditioned escalation problem, where uncertainty quantification governs the boundary between automated AI and human intervention. Current enterprise governance practices rely on categorical automation rules that conflate risk category classification with model confidence, creating a critical epistemic gap. This paper formalizes the Confidence-Aware Escalation (CAE) framework, which integrates quantified uncertainty into escalation decisions via conformal prediction theory. Conformal prediction provides distribution-free, finite-sample coverage guarantees, enabling governance policies to be specified in terms of verifiable error rate bounds rather than heuristic confidence thresholds. The CAE framework classifies AI outputs into three automation tiers by jointly evaluating prediction-set cardinality and risk category. A three-layer governance architecture comprising inference production, human oversight, and regulatory compliance is proposed, supported by structured stakeholder roles: Risk Owners, Model Stewards, and Executives. Adaptive feedback learning pipelines maintain coverage guarantees under concept drift without reward-hacking incentives. A policy-driven fairness monitoring protocol resolves mathematical incompatibility among fairness criteria through organizational policy specification rather than technical compromise. Empirical validation on a publicly available financial risk benchmark dataset confirms that the framework achieves high automation rates while maintaining near-theoretical coverage guarantees and enabling transparent governance of false escalation risk. The framework is aligned with EU AI Act high-risk classification requirements and US Federal Reserve SR 11-7 model risk management guidance.
Janardhana Naidu Kola· International journal of com...· 0 citations
Artificial intelligence (AI) is increasingly embedded in organisational systems, shaping decision making, resource management, and crisis response. Achieving AI-resilient continuity is not just about faster recovery or stronger systems; it involves maintaining decision integrity, governance quality, and human judgment during intelligent disruptions. While AI enhances efficiency and resilience, it also introduces risks that traditional continuity planning does not fully address. Unlike conventional disruptions, AI-enabled attacks may not cause immediate system failures, yet they can degrade decision accuracy, situational awareness, and governance even while operations appear normal. This paper examines how AI alters the nature of disruption and challenges assumptions about visibility, human oversight, and linear recovery. Through practical scenarios, it illustrates how data poisoning, adversarial inputs, compromised models, platform dependencies, and misinformation threaten organisational continuity. The paper proposes principles for AI-resilient planning that protect decision integrity, enable human override, strengthen governance alignment, and incorporate AI-specific exercises. This framework provides practitioners with actionable guidance for sustaining reliable decision making under AI-driven disruption. This article is also included in The Business & Management Collection which can be accessed at https:// hstalks.com/business/.
S. Haynes· Journal of Business Continui...· 0 citations
In modern IT operations (IT-Ops), the cost of an incorrect repair often exceeds the cost of no action at all. Yet existing automated remediation systems are designed to generate actions rather than to decide whether intervention is warranted, leaving safety as an afterthought enforced by manual approval. This paper makes three contributions to close this gap: (i) we reformulate safe remediation as a risk-constrained intervention decision problem and cast it as a Constrained Markov Decision Process (CMDP), in which the agent maximizes repair success subject to a bounded false remediation rate (FRR); (ii) we introduce a three-dimensional risk decomposition comprising blast radius, reversibility, and epistemic uncertainty, providing operators with an interpretable per-action safety interface; and (iii) we design a context-adaptive human-in-the-loop (HITL) gate that turns escalation from a binary failsafe into a bandwidth-aware control layer responsive to on-call load and business criticality. The full policy is learned offline from historical incident logs, enabling explicit control of the expected FRR. Experiments on the Train Ticket microservice benchmark with Chaos Mesh fault injection and an RCAEval-aligned fault taxonomy show that our framework reduces FRR by 39% while improving repair success by 2.5 points over a strong runbook baseline, and reduces on-call escalation load by 17% relative to a fixed-threshold variant.
Chengxiao Dai, Zhaokun Yan, Chenjun Lei et al.· 0 citations