Protecting Vulnerable Voices: Synthetic Dataset Generation for Self-Disclosure Detection
This work addresses privacy risks for vulnerable populations on social platforms by enabling reproducible research into self-disclosure detection, though it is incremental in applying existing LLMs to a specific data generation task.
The paper tackles the lack of open-source labeled datasets for detecting personal information disclosures in online social media by developing a novel methodology to generate synthetic PII-revealing data using large language models, resulting in a released dataset that meets reproducibility, unlinkability, and indistinguishability criteria.
Social platforms such as Reddit have a network of communities of shared interests, with a prevalence of posts and comments from which one can infer users' Personal Information Identifiers (PIIs). While such self-disclosures can lead to rewarding social interactions, they pose privacy risks and the threat of online harms. Research into the identification and retrieval of such risky self-disclosures of PIIs is hampered by the lack of open-source labeled datasets. To foster reproducible research into PII-revealing text detection, we develop a novel methodology to create synthetic equivalents of PII-revealing data that can be safely shared. Our contributions include creating a taxonomy of 19 PII-revealing categories for vulnerable populations and the creation and release of a synthetic PII-labeled multi-text span dataset generated from 3 text generation Large Language Models (LLMs), Llama2-7B, Llama3-8B, and zephyr-7b-beta, with sequential instruction prompting to resemble the original Reddit posts. The utility of our methodology to generate this synthetic dataset is evaluated with three metrics: First, we require reproducibility equivalence, i.e., results from training a model on the synthetic data should be comparable to those obtained by training the same models on the original posts. Second, we require that the synthetic data be unlinkable to the original users, through common mechanisms such as Google Search. Third, we wish to ensure that the synthetic data be indistinguishable from the original, i.e., trained humans should not be able to tell them apart. We release our dataset and code at https://netsys.surrey.ac.uk/datasets/synthetic-self-disclosure/ to foster reproducible research into PII privacy risks in online social media.