Adekunbi Justina Ajayi

This study evaluates the feasibility of post-quantum cryptographic (PQC) algorithms in mitigating "Harvest Now, Decrypt Later" (HNDL) attacks as quantum computing advances threaten classical encryption. Using datasets from the NIST Post-Quantum Cryptography Project, Deloitte’s PQC Adoption Survey, and IBM & Google Quantum Roadmaps, the study applied statistical modeling, Monte Carlo simulations, and ARIMA forecasting to assess PQC security resilience, adoption readiness, and quantum decryption feasibility. The findings indicate that CRYSTALS-Kyber and CRYSTALS-Dilithium outperform RSA-2048 and ECC-256 in quantum resistance, with attack cost thresholds exceeding 1.91×10101.91 \times 10^{10}1.91×1010 compared to 2.24×1082.24 \times 10^82.24×108 for RSA-2048. Industry adoption varies, with government PQC implementation at 79%, finance at 67%, and healthcare lagging at 48%. Quantum decryption probability remains negligible until 2029 but rises to 78.6% by 2033. This study addresses the critical challenge of post-quantum cryptographic (PQC) adoption amid growing quantum threats, particularly "Harvest Now, Decrypt Later" (HNDL) attacks. By evaluating PQC security resilience, industry adoption, and quantum decryption feasibility, the findings provide strategic insights for policymakers and cybersecurity experts. Highlighting the vulnerabilities of CRYSTALS-Kyber to side-channel attacks, this research underscores the need for continuous cryptanalysis, hybrid encryption models, and adaptive security frameworks to future-proof data.Organizations must accelerate PQC adoption, integrate hybrid cryptographic models, enforce regulatory policies, and sustain cryptanalysis efforts to ensure long-term security.

This study examines the ethical and legal implications of Shadow AI in healthcare, finance, and education by analyzing unauthorized AI deployments and their impact on data privacy, cybersecurity, and regulatory compliance. Using a quantitative research approach, descriptive statistics, ordinal regression modeling, and network analysis were employed to assess AI violations using the MITRE ATLAS AI Incident Database, EU AI Act Public Database, and IBM X-Force Threat Intelligence Report. Findings reveal that privacy breaches are most prevalent in education (22 cases), bias-related issues dominate finance (20 cases), and cybersecurity risks are highest in healthcare (19 cases). Legal risk analysis shows a 20% probability of regulatory intervention, with breach type as the strongest determinant. Anomaly detection identified healthcare as the most vulnerable to AI-driven cyber threats (8 anomalies). This study contributes to AI governance literature by quantifying the impact of regulatory interventions on Shadow AI risks, demonstrating how enforcement actions influence unauthorized AI adoption trends. It also underscores the limitations of current frameworks (e.g., GDPR, HIPAA, SEC regulations) in mitigating AI-related violations. The findings emphasize the urgent need for sector-specific AI compliance frameworks, AI ethics committees, and real-time cybersecurity monitoring systems to mitigate risks. Strengthening legal accountability and regulatory enforcement is critical to preventing the unchecked proliferation of Shadow AI in sensitive industries. Recommendations include sector-specific AI compliance frameworks, AI ethics committees, cybersecurity policies, and stricter regulatory enforcement.

The increasing adoption of cloud-based Electronic Health Records (EHRs) has transformed healthcare by enhancing data accessibility, interoperability, and patient care efficiency. However, this transition has also introduced new cybersecurity vulnerabilities, with ransomware emerging as a critical threat to healthcare systems. Ransomware attacks disrupt medical services, compromise patient confidentiality, and impose significant financial burdens on institutions. This study comprehensively examines ransomware threats in cloud-based EHR environments by analyzing vulnerabilities, attack vectors, and mitigation strategies through the NIST Cybersecurity Framework and MITRE ATT&CK Framework. A quantitative analysis was conducted using datasets from the U.S. Department of Health and Human Services (HHS), the Cybersecurity & Infrastructure Security Agency (CISA), and the MITRE ATT&CK database. Key findings indicate a 67% increase in ransomware incidents from 2018 to 2023, with credential theft (33.3%) and phishing (26.7%) as the most exploited attack vectors. Recovery challenges were exacerbated by backup failures (hazard ratio = 0.000, p = 0.127) and third-party risks (hazard ratio = 0.000, p = 0.030). To mitigate these risks, the study advocates for a multi-layered cybersecurity approach, emphasizing Zero Trust Architecture, AI-driven threat detection, immutable backups, and vendor risk management. The findings underscore the need for collaboration among healthcare institutions, cybersecurity professionals, and policymakers to strengthen resilience against evolving ransomware threats. By integrating structured cybersecurity frameworks and proactive defense mechanisms, healthcare organizations can enhance data security, ensure compliance, and minimize operational disruptions.

Cloud misconfigurations in financial Software-as-a-Service (SaaS) applications pose significant cybersecurity risks, leading to data breaches, financial losses, and reputational harm. This study utilizes data from the Cloud Security Alliance (CSA) Top Threats Dataset, Verizon Data Breach Investigations Report (DBIR), and the MITRE ATT&CK Framework to examine the causes and types of misconfigurations, analyze their financial impact, and evaluate the effectiveness of mitigation strategies. A Chi-Square Test for Independence, Ordinary Least Squares (OLS) Regression, and Kaplan-Meier Survival Analysis were employed to quantify these risks. Findings indicate that IAM errors (183 occurrences) and exposed APIs (156 occurrences) are the most frequent misconfigurations, with high-severity misconfigurations resulting in an average financial loss of $7.6M and regulatory fines of $2.5M. Implementation of Zero Trust Architecture, Cloud Security Posture Management (CSPM), and strict IAM controls reduced breach probability from 70% to 40%. This study examines cloud misconfigurations in financial SaaS applications using datasets from 2018 to 2024, providing a risk quantification of security threats and a financial impact assessment of breaches. Findings reveal that IAM errors and exposed APIs are the most frequent misconfigurations, causing severe financial losses. The effectiveness of Zero Trust and CSPM in reducing breach probability is analyzed. Future research should explore AI-driven security solutions for real-time misconfiguration detection and automated risk prevention in cloud environments. Recommendations include automating security controls, enforcing Zero Trust policies, integrating security training, and strengthening regulatory compliance.

Artificial intelligence (AI) is revolutionizing cybersecurity in digital currency transactions by enhancing fraud detection and risk mitigation. This study leverages machine learning techniques, including logistic regression and random forest classifiers, to evaluate AI’s efficacy in detecting fraudulent transactions using datasets such as the REKT Database, Elliptic Crypto Transaction Dataset, Cipher Trace AML Reports, and IEEE DataPort Financial Transactions Dataset. The analysis employs confusion matrix evaluation, fairness-aware machine learning techniques, and regression modeling to assess AI’s impact on fraud prevention. Findings reveal that AI-driven security measures reduced fraudulent activities by up to 76.86%, underscoring their effectiveness. However, the models exhibited a high false-negative rate of 89.54%, signaling a significant risk of undetected illicit transactions. Algorithmic bias is also evident, with a disparate impact ratio of 0.7793, indicating fairness concerns in AI fraud detection. These limitations highlight the need for adversarial training, fairness-aware optimization techniques, and anomaly detection refinements to improve model reliability. Additionally, regulatory frameworks, such as the EU AI Act, must be considered to ensure compliance, ethical fairness, and transparency in AI-powered cybersecurity applications. This study emphasizes the dual role of AI in strengthening fraud detection while presenting new challenges, such as algorithmic bias and adversarial vulnerabilities. To enhance AI’s effectiveness, quantum-proof encryption, fairness-aware fraud detection models, and transparent governance frameworks are recommended. These findings contribute to the ongoing discourse on AI-driven cybersecurity in digital finance, offering insights into mitigating risks while harnessing AI’s potential for securing digital currency ecosystems.

Adversarial attacks pose a critical threat to the reliability of AI-driven systems, exploiting vulnerabilities at the data, model, and deployment levels. This study employs a quantitative analysis using the CIFAR-10 Adversarial Examples Dataset from IBM’s Adversarial Robustness Toolbox and the MITRE ATLAS AI Model Vulnerabilities Dataset to assess attack success rates and attack surface exposure. A convolutional neural network (CNN) classifier was evaluated against Fast Gradient Sign Method (FGSM), Projected Gradient Descent (PGD), and Carlini & Wagner (C&W) attacks, yielding misclassification rates of 42.2%, 65.5%, and 86.8%, respectively. Statistical analysis using the Chi-Square Goodness-of-Fit Test (p < 0.001) confirmed a disproportionate targeting of model-level vulnerabilities (53.6%). These vulnerabilities pose severe risks across real-world AI applications. In cybersecurity, adversarial perturbations compromise intrusion detection systems, malware classification models, and spam filters, allowing cybercriminals to bypass AI-driven defenses. In autonomous vehicles, subtle adversarial modifications to traffic signs and road patterns can mislead AI-based navigation, increasing the likelihood of accidents. Similarly, in financial systems, adversarial attacks deceive fraud detection models, enabling unauthorized transactions and financial fraud. Countermeasure evaluation demonstrated that adversarial training provided the highest robustness gain (23.29%), while detection algorithms were least effective (15.34%). To enhance AI security, hybrid defense mechanisms integrating adversarial training with real-time anomaly detection should be prioritized, and standardized evaluation benchmarks should be established for AI security testing. These findings emphasize the necessity of hybrid AI security frameworks that combine adversarial training with real-time anomaly detection. Moreover, standardized security benchmarks should be established to ensure resilience across industries, particularly in high-stakes AI applications.