Academic Publication
Industry 4.0, Winter Session, Vol. 2, pp. 320–321
Journal: Industry 4.0 • ISSN: 2534-997X • STUME Journals
Abstract
This publication proposes a concept for implementing artificial intelligence in administrative activities, grounded in the principles of cryptographic transparency and mathematical verification. The central problem is the structural tension between the accountability requirements of Regulation (EU) 2024/1689 (AI Act) and data protection under GDPR.
The proposed solution is the ZKAP (Zero-Knowledge Audit Proxy) methodology — zero-knowledge proofs that enable mathematical proof of the legal correctness of a language model’s output, without requiring access to raw processing data.
The concept of digitalization of legal norms envisions transforming laws into arithmetic constraints embedded directly into the computational logic of the model. This achieves Safety-by-Design, making the system resistant to logical deviations.
Keywords
Artificial Intelligence, AI Act, ZKAP, Zero-Knowledge Audit Proxy, Transparency, Accountability, Administrative Activities, Cryptographic Verification.
Full Text & Related Materials
Full Text:
Zenodo DOI 10.5281/zenodo.19614243 — Open Access PDF →
ResearchGate — Full-text available →
STUME Journals — Industry 4.0 →
Extended Analysis:
AI in Public Administration — Full blog analysis (EN) →
Next Publication [2]:
Radoslavov, R.Y. (2026). “Management and Regulation of AI Models in Public Administration: Cryptographic Transparency and Digitalization of Legal Norms.” Artificial Intelligence Proceedings, ISSN 3033-2923 / 3134-1667.