Controlled Generation for Private Synthetic Text
It addresses privacy concerns for AI deployment in high-stakes domains, but appears incremental as it builds on existing de-identification and HIPS theory.
The paper tackles the problem of generating privacy-preserving synthetic text for sensitive domains like healthcare and law by proposing a method using entity-aware control codes with in-context learning or prefix tuning, achieving a strong balance between privacy and utility in experiments on legal and clinical datasets.
Text anonymization is essential for responsibly developing and deploying AI in high-stakes domains such as healthcare, social services, and law. In this work, we propose a novel methodology for privacy-preserving synthetic text generation that leverages the principles of de-identification and the Hiding In Plain Sight (HIPS) theory. Our approach introduces entity-aware control codes to guide controllable generation using either in-context learning (ICL) or prefix tuning. The ICL variant ensures privacy levels consistent with the underlying de-identification system, while the prefix tuning variant incorporates a custom masking strategy and loss function to support scalable, high-quality generation. Experiments on legal and clinical datasets demonstrate that our method achieves a strong balance between privacy protection and utility, offering a practical and effective solution for synthetic text generation in sensitive domains.