Quantum-Certified Anonymization

Irreversibility Beyond Computational Hardness

---

Daniel Mo Houshmand QDaria Quantum Research, Oslo, Norway

Abstract

We present the first data anonymization system whose irreversibility is guaranteed by the Born rule of quantum mechanics rather than by computational hardness assumptions. Every deployed anonymization tool derives its randomness from a classical PRNG; an adversary who captures the PRNG state can reconstruct every "random" value and reverse the anonymization completely. We introduce QRNG-OTP-Destroy, a protocol that replaces each personally identifiable value with a quantum-random token and irreversibly destroys the mapping. Because quantum measurement outcomes are governed by the Born rule, no deterministic seed exists, and the anonymization is information-theoretically irreversible against adversaries with unbounded computational power.

We formalize three tiers of irreversibility (computational, information-theoretic, and physics-guaranteed), prove that no classical PRNG-based method achieves the strongest tier, and prove that QRNG-OTP-Destroy does. We report on an implementation with 10 progressive anonymization levels and a multi-provider entropy architecture (Rigetti, IBM Quantum, qBraid) with automatic failover to OS entropy. We validate the implementation with 966 unit and integration tests, evaluate it on the UCI Adult dataset (32,561 records), and demonstrate production-scale quantum entropy harvesting (6.8 MB from 35 IBM Quantum jobs on 156-qubit processors).

Key Contributions

#	Contribution	Formal Result
1	Three-tier irreversibility hierarchy	Lemma 1, Theorems 1-2
2	Game-based security definition with zero adversarial advantage	Theorem 9
3	10 progressive anonymization levels (L1-L10)	Section 4
4	Multi-provider quantum entropy pool (6.8 MB real quantum entropy)	Section 5
5	GDPR Recital 26 compliance argument with auditable provenance	Section 7
6	Open-source implementation with 966 tests	Section 5.1

Figures

Fig. 1: Three-Tier Irreversibility Hierarchy Computational irreversibility breaks if P = NP. Physics-guaranteed irreversibility holds regardless of computational advances.	Fig. 3: QRNG-OTP-Destroy Protocol Data flows through PII detection, level selection (L1-L10), quantum entropy acquisition, OTP mapping, substitution, and secure destruction.
Fig. 5: Runtime and Transformation Coverage Left: mean runtime per level (1,000 rows). Right: percentage of values transformed. L10 transforms 100% of values.	Fig. 7: Guarantee Comparison Across Tools Only QRNG-OTP-Destroy provides physics-guaranteed irreversibility; all others rely on computational hardness.
Fig. 8: Privacy-Utility Spectrum Privacy-utility trade-off across 10 anonymization levels. Dashed line: maximum irreversibility achievable with classical methods.	Fig. 6: Runtime Scaling Runtime scaling from 100 to 5,000 rows for L1, L5, L8, L10. QRNG-dependent levels scale linearly with unique values.

All 14 figures available in figures/ (PDF) and images/ (PNG).

Paper Structure

Section	Title	Content
1	Introduction	HNDL threat, PRNG vulnerability, Born rule guarantee
2	Threat Model	Four adversary classes (A1-A4), game-based definitions
3	Three-Tier Framework	Computational, information-theoretic, physics-guaranteed
4	Protocol Design	QRNG-OTP-Destroy, 10 anonymization levels
5	Implementation	Multi-provider entropy, 966 tests, Zipminator SDK
6	Evaluation	UCI Adult (32,561 rows), runtime benchmarks, scaling
7	GDPR Analysis	Recital 26, provenance chain, DORA Art. 7
8	Comparison	vs. k-anonymity, l-diversity, t-closeness, DP, ARX
9	Discussion	Limitations, quantum hardware costs, future work

Building the Paper

# Prerequisites: LaTeX distribution (TeX Live 2024+)
pdflatex main.tex
pdflatex main.tex   # second pass for cross-references

The pre-compiled PDF is available at main.pdf.

Implementation

The full Zipminator SDK (anonymization engine, entropy pool, multi-provider QRNG) is open-source:

pip install zipminator[all]

from zipminator import Anonymizer

anon = Anonymizer(level=10)  # Physics-guaranteed irreversibility
result = anon.anonymize(dataframe)

Source: QDaria/zipminator | PyPI: zipminator

Related Papers

This paper is part of a three-paper series on post-quantum entropy infrastructure:

1. This paper - Quantum-Certified Anonymization
2. Unilateral WiFi CSI as a NIST-Validated Entropy Source - CSI entropy extraction + PUEK
3. Certified Heterogeneous Entropy with Algebraic Randomness Extraction - Multi-source entropy composition + ARE

Patent

Norwegian Patent Application No. 20260384, filed 24 March 2026 (Patentstyret). 15 claims covering the QRNG-OTP-Destroy protocol and quantum-certified anonymization framework.

Citation

@misc{houshmand2026quantum,
  author       = {Houshmand, Daniel Mo},
  title        = {Quantum-Certified Anonymization: Irreversibility Beyond Computational Hardness},
  year         = {2026},
  doi          = {10.5281/zenodo.19437010},
  url          = {https://doi.org/10.5281/zenodo.19437010},
}

License

Apache-2.0