Abstract
Recently, organizations supplies many ingredients of their products from third-party organizations, where security of the third-party organizations is a crucial for the success of the product. This circumstance makes, the rapid risk assessment of the organizations a challenging task. This paper introduces GEN-TPRM, a third-party security risk assessment model that uses open-source intelligence (OSINT), LLM-based question answering, and scoring aligned with ISO/IEC 27001:2022 and Gartner risk thresholds. It integrates publicly verifiable indicators like breach disclosures, CVEs, and regulatory penalties into risk-specific directional QA replies and a standardized risk index. The model integrates a retrieval-augmented LLM layer with a logistic regression classifier trained on historical security cases. The approach was validated through a real-world case study of 100 well-known vendors. The model generates transparent outputs without requiring internal access to vendor environments, offering a lightweight, scalable solution for lifecycle-based third-party pre-risk management.
| Original language | English |
|---|---|
| Title of host publication | 2025 15th International Conference on Advanced Computer Information Technologies, ACIT 2025 - Conference Proceedings |
| Publisher | Institute of Electrical and Electronics Engineers |
| Pages | 497-500 |
| Number of pages | 4 |
| ISBN (Electronic) | 9798331595432 |
| DOIs | |
| Publication status | Published - 2025 |
| Event | 15th International Conference on Advanced Computer Information Technologies, ACIT 2025 - Hybrid, Sibenik, Croatia Duration: 17 Sept 2025 → 19 Sept 2025 |
Publication series
| Name | Proceedings - International Conference on Advanced Computer Information Technologies, ACIT |
|---|---|
| ISSN (Print) | 2770-5218 |
| ISSN (Electronic) | 2770-5226 |
Conference
| Conference | 15th International Conference on Advanced Computer Information Technologies, ACIT 2025 |
|---|---|
| Country/Territory | Croatia |
| City | Hybrid, Sibenik |
| Period | 17/09/25 → 19/09/25 |
Bibliographical note
Publisher Copyright:© 2025 IEEE.
Keywords
- augmented generation
- large language model
- security
- source intelligence
- TPRM