Kateryna Shcheglova
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Kateryna Shcheglova has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Liubomyr Kochubei
Master of Law, Master of Public Administration, Attorney, Chairman of the IT Lawyers Bar Association, Ukraine
2 shared publications - Oleksandr Levchenko
Ph.D. in Economics, Expert in Financial Regulation and Prudential Supervision, Certified Bank Auditor, Ukraine
2 shared publications - Oleksiy Domashenko
MBA, EA, Registered Investment Advisor, Accounting and Tax Expert, USA
2 shared publications - Zhanna Dryha
Doctor of Economics, Professor, Department of National Security, Public Administration and Administration, Zhytomyr Polytechnic State University, Ukraine
2 shared publications Anton Pylypets
1 shared publication- Larysa Sergiienko
Zhytomyr Polytechnic State University, Faculty of National Security, Law and International Relations, Department of National Security, Public Management and Administration, Zhytomyr, Ukraine
1 shared publication
Latest Publications
- Open Access
- 9 October 2026
- Published by European Scientific Platform (Publications)
- No. October 9,,p. 8-15
AI ASSURANCE AS AN ARCHITECTURE OF TRUST IN PUBLIC FINANCIAL REGULATION IN THE AGENTIC AI ERA
Artificial intelligence is changing both the object and the material form of public financial supervision. Traditional Finance, FinTech, Distributed Finance and Agentic Finance increasingly coexist, while digital modularity, programmability, distributed coordination and delegated autonomous action expand the regulatory object from an institution or model to an ecosystem of people, regulated firms, AI agents, models, data, APIs, infrastructure and supervisory institutions [Financial Conduct Authority, 2026a; Davidovic & Tourpe, 2026]. Existing trust mechanisms remain necessary: supervision, governance, audit, certification, operational resilience, cybersecurity, model-risk management and consumer protection. Their limitation is increasingly organisational: mechanisms developed for different risks and time cycles generate separate evidence that must now be reconciled across a continuous socio-technical system. This article therefore examines AI Assurance as an architectural mode of organising trust mechanisms when financial activity becomes agentic, autonomous, machine - readable and machine-executable.
- 9 October 2026
- Published by European Scientific Platform (Publications)
- No. October 9,,p. 8-15
Open AccessAI ASSURANCE AS AN ARCHITECTURE OF TRUST IN PUBLIC FINANCIAL REGULATION IN THE AGENTIC AI ERA
Artificial intelligence is changing both the object and the material form of public financial supervision. Traditional Finance, FinTech, Distributed Finance and Agentic Finance increasingly coexist, while digital modularity, programmability, distributed coordination and delegated autonomous action expand the regulatory object from an institution or model to an ecosystem of people, regulated firms, AI agents, models, data, APIs, infrastructure and supervisory institutions [Financial Conduct Authority, 2026a; Davidovic & Tourpe, 2026]. Existing trust mechanisms remain necessary: supervision, governance, audit, certification, operational resilience, cybersecurity, model-risk management and consumer protection. Their limitation is increasingly organisational: mechanisms developed for different risks and time cycles generate separate evidence that must now be reconciled across a continuous socio-technical system. This article therefore examines AI Assurance as an architectural mode of organising trust mechanisms when financial activity becomes agentic, autonomous, machine - readable and machine-executable.
- Open Access
- 31 August 2026
- Published by FinTechAlliance in Financial and credit activity: problems of theory and practice
- Vol. 4 (69) , 81-102
FINANCIAL, DIGITAL AND LEGAL MECHANISMS FOR REGULATORY COMPLIANCE IN NON-BANK FINANCIAL INSTITUTIONS
Ukraine's transition to risk-based prudential supervision requires non-bank financial institutions to connect formal regulatory obligations with measurable financial resilience, reliable data, governance, and professional competence. The purpose of this study is to assess the financial transformation and prudential compliance of selected Ukrainian non-bank financial institution segments, systematize applicable regulatory requirements, and develop an integrated framework linking financial, digital, legal, governance, and professional mechanisms. The analysis combines official National Bank of Ukraine data for insurance companies, finance companies, and credit unions, descriptive period comparisons, author-verified regulatory coding, and comparative analysis of continuing professional development frameworks. The results reveal market consolidation across all three segments but divergent balance-sheet trajectories: insurers and finance companies expanded their aggregate assets, whereas credit unions contracted. Governance and legal obligations dominate the primary-code structure of the selected regulatory requirements, while multiple coding shows that financial, digital, legal, governance, and professional components are interdependent. The study develops a financial-first conceptual framework in which prudential assessment forms the evidence-based core, digital tools support monitoring and traceability, legal mechanisms translate external requirements into internal obligations, and governance and professional competence operate as cross-cutting layers. The findings support differentiated implementation rather than a uniform compliance template for all non-bank financial institutions. The results may assist the National Bank of Ukraine, non-bank financial institutions, compliance, risk-management and internal-audit functions, professional associations, researchers, and policy specialists in designing proportionate control, reporting, governance, and professional-development systems.
- 31 August 2026
- Published by FinTechAlliance in Financial and credit activity: problems of theory and practice
- Vol. 4 (69) , 81-102
Open AccessFINANCIAL, DIGITAL AND LEGAL MECHANISMS FOR REGULATORY COMPLIANCE IN NON-BANK FINANCIAL INSTITUTIONS
Ukraine's transition to risk-based prudential supervision requires non-bank financial institutions to connect formal regulatory obligations with measurable financial resilience, reliable data, governance, and professional competence. The purpose of this study is to assess the financial transformation and prudential compliance of selected Ukrainian non-bank financial institution segments, systematize applicable regulatory requirements, and develop an integrated framework linking financial, digital, legal, governance, and professional mechanisms. The analysis combines official National Bank of Ukraine data for insurance companies, finance companies, and credit unions, descriptive period comparisons, author-verified regulatory coding, and comparative analysis of continuing professional development frameworks. The results reveal market consolidation across all three segments but divergent balance-sheet trajectories: insurers and finance companies expanded their aggregate assets, whereas credit unions contracted. Governance and legal obligations dominate the primary-code structure of the selected regulatory requirements, while multiple coding shows that financial, digital, legal, governance, and professional components are interdependent. The study develops a financial-first conceptual framework in which prudential assessment forms the evidence-based core, digital tools support monitoring and traceability, legal mechanisms translate external requirements into internal obligations, and governance and professional competence operate as cross-cutting layers. The findings support differentiated implementation rather than a uniform compliance template for all non-bank financial institutions. The results may assist the National Bank of Ukraine, non-bank financial institutions, compliance, risk-management and internal-audit functions, professional associations, researchers, and policy specialists in designing proportionate control, reporting, governance, and professional-development systems.
- Open Access
- 13 August 2026
- Published by LLC CPC Business Perspectives in Investment Management and Financial Innovations
- Vol. 23 (3) , 215–234-234
Explainable AI for delinquency risk monitoring in U.S. auto lease securitizations
Type of the article: Research ArticleAbstractThis study evaluates whether explainable machine-learning models can provide reliable and operationally interpretable short-horizon delinquency monitoring in U.S. auto lease securitization panels. Six public SEC ABS-EE trust-family panels were harmonized at the contract-month level. The primary outcome is one-month-ahead incident escalation to 30 or more days past due among contracts below 30 days past due at the feature month. Fully tuned penalized logistic regression (M1), unconstrained gradient boosting (M2), and governance-constrained gradient boosting (M3/X-LEASE) were assessed through chronological development, calibration, locked out-of-time testing, contract-held-out validation, six leave-one-issuer-out experiments, and later temporal evaluation. The locked test comprised 198,301 observations and 768 events. M2 produced the strongest discrimination (AUC-ROC 0.8151; PR-AUC 0.0269), while M3 exceeded the logistic benchmark by PR-AUC but did not outperform M2. At an exact 1% review capacity, each model generated 1,984 alerts; M2 detected 104 events, compared with 61 for M3, so the hypothesized recall advantage of X-LEASE was not supported. Full-sample TreeSHAP analysis showed stable feature rankings across tested issuers, later periods, and contract-level resampling. M3 relied more heavily on credit score and issuer controls and had a more concentrated explanation structure, but this does not establish superior auditability or fairness. The results support human-reviewed early-warning monitoring within the tested securitization panels, not lifetime default prediction, automated adverse decisions, or universal transferability.AcknowledgmentsThe authors acknowledge the public availability of SEC EDGAR Form ABS-EE, Exhibit 102 asset-level data. No individuals or institutions outside the author team provided paid analytical, editorial, or funding support for this manuscript.
- 13 August 2026
- Published by LLC CPC Business Perspectives in Investment Management and Financial Innovations
- Vol. 23 (3) , 215–234-234
Open AccessExplainable AI for delinquency risk monitoring in U.S. auto lease securitizations
Type of the article: Research ArticleAbstractThis study evaluates whether explainable machine-learning models can provide reliable and operationally interpretable short-horizon delinquency monitoring in U.S. auto lease securitization panels. Six public SEC ABS-EE trust-family panels were harmonized at the contract-month level. The primary outcome is one-month-ahead incident escalation to 30 or more days past due among contracts below 30 days past due at the feature month. Fully tuned penalized logistic regression (M1), unconstrained gradient boosting (M2), and governance-constrained gradient boosting (M3/X-LEASE) were assessed through chronological development, calibration, locked out-of-time testing, contract-held-out validation, six leave-one-issuer-out experiments, and later temporal evaluation. The locked test comprised 198,301 observations and 768 events. M2 produced the strongest discrimination (AUC-ROC 0.8151; PR-AUC 0.0269), while M3 exceeded the logistic benchmark by PR-AUC but did not outperform M2. At an exact 1% review capacity, each model generated 1,984 alerts; M2 detected 104 events, compared with 61 for M3, so the hypothesized recall advantage of X-LEASE was not supported. Full-sample TreeSHAP analysis showed stable feature rankings across tested issuers, later periods, and contract-level resampling. M3 relied more heavily on credit score and issuer controls and had a more concentrated explanation structure, but this does not establish superior auditability or fairness. The results support human-reviewed early-warning monitoring within the tested securitization panels, not lifetime default prediction, automated adverse decisions, or universal transferability.AcknowledgmentsThe authors acknowledge the public availability of SEC EDGAR Form ABS-EE, Exhibit 102 asset-level data. No individuals or institutions outside the author team provided paid analytical, editorial, or funding support for this manuscript.
- Open Access
- 12 February 2026
- Published by Zhytomyr Polytechnic State University in Public Policy and Accounting
- No. 1(13),p. 41-49
Invariance-based AI governance methodology through a distributed risk consensus mechanism in agentic and autonomous financial ecosystems
Human-in-the-loop was a relevant governance model for earlier AI systems because they mainly produced recommendations or alerts at a volume and speed that allowed people to review consequential outputs before execution. Agentic AI changes this premise. Autonomous agents can plan, act, interact and multiply decisions, dependencies and risks at machine speed, making action-by-action human review structurally infeasible. Purpose. The article develops a conceptual invariance-based AI governance methodology that strengthens meaningful human oversight through an intelligent supervising layer of Sentinel or guardian agents with embedded distributed risk consensus mechanism. Results. An invariant is treated as a stable condition, boundary or state that must remain preserved regardless of the trajectory by which an autonomous system reaches an operational objective. Specialised financial, compliance, cybersecurity, operational-resilience, conduct and ethics Sentinel agents assess domain invariants independently and in parallel. The distributed risk consensus engine aggregates domain-specific Sentinel signals and forms a consensus invariant status: preserved, at risk, uncertain, violated or critically violated. Humans remain at the upper governance level, where they define invariant libraries, thresholds, weights, escalation rules, override rights and suspension conditions. Conclusion. The proposed architecture transforms human oversight from control of individual actions into governance of the boundaries of permissible autonomy.
- 12 February 2026
- Published by Zhytomyr Polytechnic State University in Public Policy and Accounting
- No. 1(13),p. 41-49
Open AccessInvariance-based AI governance methodology through a distributed risk consensus mechanism in agentic and autonomous financial ecosystems
Human-in-the-loop was a relevant governance model for earlier AI systems because they mainly produced recommendations or alerts at a volume and speed that allowed people to review consequential outputs before execution. Agentic AI changes this premise. Autonomous agents can plan, act, interact and multiply decisions, dependencies and risks at machine speed, making action-by-action human review structurally infeasible. Purpose. The article develops a conceptual invariance-based AI governance methodology that strengthens meaningful human oversight through an intelligent supervising layer of Sentinel or guardian agents with embedded distributed risk consensus mechanism. Results. An invariant is treated as a stable condition, boundary or state that must remain preserved regardless of the trajectory by which an autonomous system reaches an operational objective. Specialised financial, compliance, cybersecurity, operational-resilience, conduct and ethics Sentinel agents assess domain invariants independently and in parallel. The distributed risk consensus engine aggregates domain-specific Sentinel signals and forms a consensus invariant status: preserved, at risk, uncertain, violated or critically violated. Humans remain at the upper governance level, where they define invariant libraries, thresholds, weights, escalation rules, override rights and suspension conditions. Conclusion. The proposed architecture transforms human oversight from control of individual actions into governance of the boundaries of permissible autonomy.