Security & Future
Vol. 10 (2026), Issue 1, pg(s) 20-25

INFORMATION SECURITY

Zkap anti-corruption output filter: a cryptographically verifiable compliance protocol for the deployment of artificial intelligence in investigative, regulatory, supervisory and judicial procedures

  • 1 Independent researchers Law Office — Ruse, Bulgaria • Advanced Consulting London RR, United Kingdom

Abstract

The integration of artificial intelligence into judicial and administrative procedures raises a structural regulatory crisis. The European Union has imposed simultaneous and competing obligations: on one hand, the AI Act (Regulation (EU) 2024/1689) mandates transparency, human oversight, and accuracy for high-risk systems; on the other, the GDPR (Regulation (EU) 2016/679) and the Trade Secrets Directive (2016/943) impose strict confidentiality. This article proposes the ZKAP (Zero-Knowledge Audit Protocol) Anti-Corruption Output Filter as a cryptographically verifiable compliance mechanism that resolves the conflict at the operational layer of AI-assisted judicial workflows. The filter operates on a universal dual-axis decomposition of any procedural norm into a fixed set of mandatory positions (X) and conditions (Y); output generation is automatically blocked when any element is unmet, and any manual override triggers a non-disclosable escalation signal to the institutional supervisor, containing only metadata. We demonstrate the architecture through a polynomial decomposition of one specific Y-element — the determination of the circle of investigated persons under the Bulgarian Asset Recovery Act (ZONPI), implementing § 1, items 6, 15, 18 and 19 of its Supplementary Provisions, Articles 6, 7 and 143 of the Family Code and Article 2 of the Persons and Family Act. The polynomial is validated on a fully synthetic test family configuration designed to exercise all nine qualification members; the synthetic profile is purely illustrative and contains no personal or case-related data. The chosen example is intentionally one of the architecturally shallower components of the filter, used here to illustrate the polynomialization principle; in parallel deployment, substantively deeper components — banking pattern analysis, real estate transformation, fiscal coherence, and the determination of significant asset disparity — are similarly polynomialized but remain outside the scope of this article. The architecture is patent-protected under Bulgarian filings 114317 and 114328, with EPO/PCT/UKIPO submissions in preparation. The contribution proposes the architecture as an operational compliance mechanism that can be implemented either through institutional codes of conduct or through normative acts, depending on the procedure-specific corruption-risk profile.

Keywords

References

  1. A. Normative acts and official EU publications
  2. Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 (General Data Protection Regulation). OJ L 119, 4.5.2016, p. 1–88. CELEX: 32016R0679.
  3. Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 (Artificial Intelligence Act). OJ L, 2024/1689, 12.7.2024. CELEX: 32024R1689.
  4. Directive (EU) 2016/943 of the European Parliament and of the Council of 8 June 2016 on the protection of undisclosed know-how and business information (trade secrets). OJ L 157, 15.6.2016, p. 1–18. CELEX: 32016L0943.
  5. B. Case law
  6. Court of Justice of the European Union (2023). Judgment of 7 December 2023, OQ v Land Hessen (SCHUFA Holding), Case C-634/21, ECLI:EU:C:2023:957.
  7. Supreme Court of Cassation of the Republic of Bulgaria, General Assembly of the Civil Chamber (2023). Interpretative Decision No. 4 of 18 May 2023 in Interpretative Case No. 4/2021.
  8. C. National legislation of the Republic of Bulgaria
  9. Act on the Forfeiture of Illegally Acquired Property (ZONPI; promulgated as ZPKONPI and retitled by SG No. 84 of 2023). State Gazette No. 7 of 19 January 2018; last amended SG No. 61 of 2025 (in force from 31 January 2026).
  10. Family Code of the Republic of Bulgaria. State Gazette No. 47 of 23 June 2009, subsequent amendments.
  11. Persons and Family Act of the Republic of Bulgaria. State Gazette No. 182 of 9 August 1949, subsequent amendments.
  12. D. Cryptographic and technical foundations
  13. Goldwasser, S., Micali, S., & Rackoff, C. (1989). The Knowledge Complexity of Interactive Proof Systems. SIAM Journal on Computing, 18(1), 186–208.
  14. Groth, J. (2016). On the Size of Pairing-based Non-Interactive Arguments. EUROCRYPT 2016, LNCS 9666, Springer, pp. 305–326.
  15. Ben-Sasson, E., Bentov, I., Horesh, Y., & Riabzev, M. (2018). Scalable, transparent, and post-quantum secure computational integrity. IACR ePrint 2018/046.
  16. Costello, C., et al. (2015). Geppetto: Versatile Verifiable Computation. IEEE S&P 2015, pp. 253–270.
  17. Pinto, S., & Santos, N. (2019). Demystifying Arm TrustZone. ACM Computing Surveys, 51(6), Art. 130.
  18. E. Explainable AI (XAI)
  19. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). “Why Should I Trust You?”: Explaining the Predictions of Any Classifier. KDD 2016, pp. 1135–1144.
  20. Lundberg, S. M., & Lee, S.-I. (2017). A Unified Approach to Interpreting Model Predictions. NeurIPS 2017, pp. 4765–4774.
  21. F. Legal foundations of algorithmic adjudication
  22. Kroll, J. A., et al. (2017). Accountable Algorithms. University of Pennsylvania Law Review, 165(3), 633–705.
  23. Pasquale, F. (2015). The Black Box Society. Harvard University Press.
  24. Selbst, A. D., & Powles, J. (2017). Meaningful Information and the Right to Explanation. IDPL, 7(4), 233–242.
  25. G. Own publications, patents and sources
  26. Radoslavov, R. Y. (2025). Management and Regulation of Artificial Intelligence Models: A Concept for Transparency and Accountability in Administrative Activities. Industry 4.0, Winter Session 2025, Volume 2, pp. 320–321. Sofia: STUME.
  27. Radoslavov, R. Y. (2026). ZKAP: Zero-Knowledge Audit Protocol — Professional White Paper for Cryptographic Regulation of Artificial Intelligence. Version 5.3, April 2026. London: Advanced Consulting London RR.
  28. Radoslavov, R. Y., & Genov, M. (2026). Beyond Transparency: How to Govern Artificial Intelligence Through Cryptographic Regulation in the Era of Opaque Algorithms. London: Advanced Consulting London RR.
  29. Patent application BG/P/2026/114317. Method and system for cryptographically bound inference execution in machine learning models with automatic generation of zero-knowledge proofs over formalized constraints (ZKAP). Filed 30.03.2026. IPC: G06F 21/64; G06N 20/00; H04L 9/32. Applicant: Radoslav Y. Radoslavov.
  30. Patent application BG/P/2026/114328. Method and system for a software implementation of a zero-knowledge audit protocol over formalized constraints, with asymmetric periodic attestation and sandbox-level prove-before-output enforcement. Filed 12.04.2026. IPC: G06F 21/64; G06N 20/00; H04L 9/32. Applicant: Radoslav Y. Radoslavov. EPO/PCT/UKIPO submissions in preparation. FRAND-ready declaration available.
  31. Radoslavov, R. Y. (2026). ZKAP — Zero-Knowledge Audit Protocol — Interactive Demonstrator. Hash-based simulation prototype; production implementation extends to SNARK verification and TEE/Silicon Binding.

Article full text

Download PDF