Measuring Privacy vs. Fidelity in Synthetic Social Media Datasets

TL;DR

This paper evaluates the privacy risks and data fidelity of synthetic social media posts generated by large language models, proposing a framework to quantify privacy leakage and analyze the privacy-fidelity trade-off.

Contribution

It introduces a novel methodology for assessing privacy in synthetic text data using authorship attribution attacks and examines the privacy-fidelity balance in social media datasets.

Findings

Synthetic posts show reduced re-identification risk compared to real data.

Higher fidelity in synthetic data correlates with increased privacy leakage.

The proposed framework effectively quantifies privacy risks in synthetic social media text.

Abstract

Synthetic data is increasingly used to support research without exposing sensitive user content. Social media data is one of the types of datasets that would hugely benefit from representative synthetic equivalents that can be used to bootstrap research and allow reproducibility through data sharing. However, recent studies show that (tabular) synthetic data is not inherently privacy-preserving. Much less is known, however, about the privacy risks of synthetically generated unstructured texts. This work evaluates the privacy of synthetic Instagram posts generated by three state-of-the-art large language models using two prompting strategies. We propose a methodology that quantifies privacy by framing re-identification as an authorship attribution attack. A RoBERTa-large classifier trained on real posts achieved 81\% accuracy in authorship attribution on real data, but only 16.5--29.7\% on synthetic posts, showing reduced, though non-negligible, risk. Fidelity was assessed via text traits, sentiment, topic overlap, and embedding similarity, confirming the expected trade-off: higher fidelity coincides with greater privacy leakage. This work provides a framework for evaluating privacy in synthetic text and demonstrates the privacy--fidelity tension in social media datasets.