mlsanityData to 7 Oct 2026

Synthetic Data -- Anonymisation Groundhog Day

arXiv 2011.07018 ↗First posted 13 Nov 2020cs.LG cs.CRcited by 105

Not on the map. Why?

Abstract (arXiv)

Synthetic data has been advertised as a silver-bullet solution to privacy-preserving data publishing that addresses the shortcomings of traditional anonymisation techniques. The promise is that synthetic data drawn from generative models preserves the statistical properties of the original dataset but, at the same time, provides perfect protection against privacy attacks. In this work, we present the first quantitative evaluation of the privacy gain of synthetic data publishing and compare it to that of previous anonymisation techniques. Our evaluation of a wide range of state-of-the-art generative models demonstrates that synthetic data either does not prevent inference attacks or does not retain data utility. In other words, we empirically show that synthetic data does not provide a better tradeoff between privacy and utility than traditional anonymisation techniques. Furthermore, in contrast to traditional anonymisation, the privacy-utility tradeoff of synthetic data publishing is hard to predict. Because it is impossible to predict what signals a synthetic dataset will preserve and what information will be lost, synthetic data leads to a highly variable privacy gain and unpredictable utility loss. In summary, we find that synthetic data is far from the holy grail of privacy-preserving data publishing.

Key facts. Cited by 105 papers in the map (data through 7 Oct 2026). First posted on arXiv 13 Nov 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 3; 2022: 12; 2023: 28; 2024: 25; 2025: 15; 2026: 22.

105 papers.

Sort:

By area:

  1. 2023Machine Learning for Synthetic Data Generation: A Reviewcited by 118
  2. 2023Differentially Private Diffusion Models Generate Useful Synthetic Imagescited by 85
  3. 2023Membership Inference Attacks against Diffusion Modelscited by 84
  4. 2022Synthetic Data -- what, why and how?cited by 66
  5. 2022TAPAS: a Toolbox for Adversarial Privacy Auditing of Synthetic Datacited by 56
  6. 2024Fairness Feedback Loops: Training on Synthetic Data Amplifies Biascited by 55
  7. 2024SoK: Membership Inference Attacks on LLMs are Rushing Nowhere (and How to Fix It)cited by 54
  8. 2022Privacy-friendly Synthetic Data for the Development of Face Morphing Attack Detectorscited by 47
  9. 2022SoK: Let the Privacy Games Begin! A Unified Treatment of Data Inference Privacy in Machine Learningcited by 47
  10. 2022Adversarial random forests for density estimation and generative modelingcited by 44
  11. 2023SoK: Privacy-Preserving Data Synthesiscited by 37
  12. 2023Beyond Privacy: Navigating the Opportunities and Challenges of Synthetic Datacited by 29
38 more
  1. 2023Achilles' Heels: Vulnerable Record Identification in Synthetic Data Publishingcited by 28
  2. 2024ClavaDDPM: Multi-relational Data Synthesis with Cluster-guided Diffusion Modelscited by 28
  3. 2022Generating multivariate time series with COmmon Source CoordInated GAN (COSCI-GAN)cited by 23
  4. 2023A Linear Reconstruction Approach for Attribute Inference Attacks against Synthetic Datacited by 23
  5. 2024Nearly Tight Black-Box Auditing of Differentially Private Machine Learningcited by 19
  6. 2023Synthetic is all you need: removing the auxiliary data assumption for membership inference attacks against synthetic datacited by 18
  7. 2025The Canary's Echo: Auditing Privacy Risks of LLM-Generated Synthetic Textcited by 18
  8. 2023Membership Inference Attacks on Diffusion Models via Quantile Regressioncited by 17
  9. 2024Systematic Assessment of Tabular Data Synthesiscited by 17
  10. 2021Data synthesis and adversarial networks: A review and meta-analysis in cancer imagingcited by 15
  11. 2025A False Sense of Privacy: Evaluating Textual Data Sanitization Beyond Surface-level Privacy Leakagecited by 15
  12. 2022medigan: a Python library of pretrained generative models for medical image synthesiscited by 14
  13. 2023CuTS: Customizable Tabular Synthetic Data Generationcited by 12
  14. 2023Recent Advances of Differential Privacy in Centralized Deep Learning: A Systematic Surveycited by 12
  15. 2024Establishing and Evaluating Trustworthy AI: Overview and Research Challengescited by 12
  16. 2023Synthetic Data: Methods, Use Cases, and Riskscited by 10
  17. 2023You Don't Have to Be Perfect to Be Amazing: Unveil the Utility of Synthetic Imagescited by 10
  18. 2024Privacy Vulnerabilities in Marginals-based Synthetic Datacited by 10
  19. 2024Private Synthetic Text Generation with Diffusion Modelscited by 10
  20. 2022Non-Imaging Medical Data Synthesis for Trustworthy AI: A Comprehensive Surveycited by 9
  21. 2023Investigating the Effect of Misalignment on Membership Privacy in the White-box Settingcited by 9
  22. 2022On the Utility Recovery Incapability of Neural Net-based Differential Private Tabular Training Data Synthesizer under Privacy Deregulationcited by 8
  23. 2025A Consensus Privacy Metrics Framework for Synthetic Datacited by 8
  24. 2025Synthetic Tabular Data: Methods, Attacks and Defensescited by 8
  25. 2024Downstream Task-Oriented Generative Model Selections on Synthetic Data Training for Fraud Detection Modelscited by 7
  26. 2025Synthetic Data Privacy Metricscited by 7
  27. 2022QuerySnout: Automating the Discovery of Attribute Inference Attacks against Query-Based Systemscited by 6
  28. 2023FLAIM: AIM-based Synthetic Data Generation in the Federated Settingcited by 6
  29. 2024Synthetic Data: Revisiting the Privacy-Utility Trade-offcited by 6
  30. 2024Tabular Data Synthesis with Differential Privacy: A Surveycited by 6
  31. 2023The Real Deal Behind the Artificial Appeal: Inferential Utility of Tabular Synthetic Datacited by 5
  32. 2024Generating Synthetic Health Sensor Data for Privacy-Preserving Wearable Stress Detectioncited by 5
  33. 2024DP-TLDM: Differentially Private Tabular Latent Diffusion Modelcited by 5
  34. 2024DP-2Stage: Adapting Language Models as Differentially Private Tabular Data Generatorscited by 5
  35. 2025Trustworthy AI: Safety, Bias, and Privacy -- A Surveycited by 5
  36. 2025Privacy Auditing Synthetic Data Release through Local Likelihood Attackscited by 5
  37. 2022Provable Membership Inference Privacycited by 4
  38. 2022Differentially-Private Data Synthetisation for Efficient Re-Identification Risk Controlcited by 4

50 most cited of 105 All 105 (JSON)