DSpace Repository

From Threat Intelligence to Economic Metrics: LLM-supported Cyber Riskops for Continuous SV/ROSI Assessment

Show simple item record

dc.contributor.author Belinschi, Ghenadie
dc.contributor.author Dorogaia, Irina
dc.date.accessioned 2026-11-13T09:27:06Z
dc.date.available 2026-11-13T09:27:06Z
dc.date.issued 2026
dc.identifier.isbn 978-9975-182-48-5 (PDF)
dc.identifier.uri https://irek.ase.md:443/xmlui/handle/123456789/5618
dc.description BELINSCHI, Ghenadie and Irina DOROGAIA. From Threat Intelligence to Economic Metrics: LLM-supported Cyber Riskops for Continuous SV/ROSI Assessment. Online. In: Technological Innovations and Digital Security: Proceedings of Scientific Articles Presented at the International Conference, 2nd Edition, Chisinau, May 14-15, 2026. Chișinău: SEP ASEM, 2026, pp. 224-232. ISBN 978-9975-182-48-5 (PDF). Disponibil: https://doi.org/10.53486/tids2026.27 en_US
dc.description.abstract This paper examines the gap between rapidly evolving cyber threats and the relatively static economic justification of cybersecurity innovations in organizations. Building on a risk -based view of information management and on the concept of Cyber RiskOps as a continuous risk-management cycle, we propose an LLM-supported pipeline that translates Threat Intelligence (TI) into auditable updates of scenario parameters used in economic metrics. The pipeline combines an information system passport—a structured register of information assets, business processes, dependencies, and controls—a scenario library, and external/internal signals. The large language model (LLM) is used strictly as an extraction and normalization component, not as an autonomous decision-making mechanism. It transforms unstructured TI into structured attributes, including actor, TTPs, affected technologies, exploitation maturity, and business-impact drivers, and maps them to scenarios through transparent scoring rules and human verification. The updated probabilities and loss magnitudes serve as inputs for the Saved Value of Information (SV) metric, defined as the avoided losses related to confidentiality, integrity, and availability, and for its linkage with Return on Security Investment (ROSI) when comparing security controls under an as-a-service subscription model. An illustrative SME case shows how faster parameter updates can change investment decisions: a package of controls that appears unprofitable under outdated assumptions may become ROSI-positive once TI-driven probability shifts are incorporated in a timely manner. The paper discusses governance requirements, including prompt and data control, audit trails, and segregation of duties, as well as limitations of the approach, such as uncertainty in loss estimation and LLM-specific risks, including hallucinations and prompt injection. The proposed approach supports doctoral research on improving information management through cybersecurity innovations by operationalizing continuous, evidence-based measurement of the economic value of security. CZU: 004.056:005.342; JEL: M15; D81; O33 en_US
dc.language.iso en en_US
dc.publisher SEP ASEM en_US
dc.subject LLM en_US
dc.subject threat intelligence en_US
dc.subject cyber risk quantification en_US
dc.subject ROSI en_US
dc.subject cybersecurity innovations en_US
dc.subject information management en_US
dc.title From Threat Intelligence to Economic Metrics: LLM-supported Cyber Riskops for Continuous SV/ROSI Assessment en_US
dc.type Article en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account