ML Security PapersDP-MGTD: Privacy-Preserving Machine-Generated Text Detection via Adaptive Differentially Private Entity Sanitization
Lionel Z. Wang 1,2, Yusheng Zhao 1,3, Jiabin Luo 1,4, Xinfeng Li 1, Lixu Wang 1, Yinan Peng 5, Haoyang Li 2, XiaoFeng Wang 1, Wei Dong 1
1 Nanyang Technological University
2 The Hong Kong Polytechnic University
Published on arXiv
2601.04641
Output Integrity Attack
OWASP ML Top 10 — ML09
Key Finding
Achieves near-perfect machine-generated text detection accuracy under strict differential privacy guarantees, outperforming non-private baselines, with DP noise counter-intuitively amplifying human-vs-machine distinguishability.
DP-MGTD
Novel technique introduced
The deployment of Machine-Generated Text (MGT) detection systems necessitates processing sensitive user data, creating a fundamental conflict between authorship verification and privacy preservation. Standard anonymization techniques often disrupt linguistic fluency, while rigorous Differential Privacy (DP) mechanisms typically degrade the statistical signals required for accurate detection. To resolve this dilemma, we propose \textbf{DP-MGTD}, a framework incorporating an Adaptive Differentially Private Entity Sanitization algorithm. Our approach utilizes a two-stage mechanism that performs noisy frequency estimation and dynamically calibrates privacy budgets, applying Laplace and Exponential mechanisms to numerical and textual entities respectively. Crucially, we identify a counter-intuitive phenomenon where the application of DP noise amplifies the distinguishability between human and machine text by exposing distinct sensitivity patterns to perturbation. Extensive experiments on the MGTBench-2.0 dataset show that our method achieves near-perfect detection accuracy, significantly outperforming non-private baselines while satisfying strict privacy guarantees.
Key Contributions
- Adaptive Differentially Private Entity Sanitization algorithm with a two-stage mechanism (noisy frequency estimation + dynamic privacy budget calibration) for privacy-preserving MGT detection
- Counter-intuitive empirical finding that DP noise amplifies distinguishability between human and machine text by exposing differential sensitivity patterns
- Near-perfect MGT detection accuracy on MGTBench-2.0 while satisfying strict DP guarantees, outperforming non-private baselines
🛡️ Threat Analysis
The primary contribution is a novel machine-generated text detection framework (DP-MGTD). AI-generated content detection maps directly to ML09 (output integrity and content authenticity). The paper proposes a new detection architecture rather than merely applying an existing detector to a new domain.