Research

Paper

AI LLM March 20, 2026

Text-Based Personas for Simulating User Privacy Decisions

Authors

Kassem Fawaz, Ren Yi, Octavian Suciu, Rishabh Khandelwal, Hamza Harkous, Nina Taft, Marco Gruteser

Abstract

The ability to simulate human privacy decisions has significant implications for aligning autonomous agents with individual intent and conducting cost-effective, large-scale privacy-centric user studies. Prior approaches prompt Large Language Models (LLMs) with natural language user statements, data-sharing histories, or demographic attributes to simulate privacy decisions. These approaches, however, fail to balance individual-level accuracy, prompt usability, token efficiency, and population-level representation. We present Narriva, an approach that generates text-based synthetic privacy personas to address these shortcomings. Narriva grounds persona generation in prior user privacy decisions, such as those from large-scale survey datasets, rather than purely relying on demographic stereotypes. It compresses this data into concise, human-readable summaries structured by established privacy theories. Through benchmarking across five diverse datasets, we analyze the characteristics of Narriva's synthetic personas in modeling both individual and population-level privacy preferences. We find that grounding personas in past privacy behaviors achieves up to 88% predictive accuracy (significantly outperforming a non-personalized LLM baseline), and yields an 80-95% reduction in prompt tokens compared to in-context learning with raw examples. Finally, we demonstrate that personas synthesized from a single survey can reproduce the aggregate privacy behaviors and statistical distributions (TVComplement up to 0.85) of entirely different studies.

Metadata

arXiv ID: 2603.19791
Provider: ARXIV
Primary Category: cs.CR
Published: 2026-03-20
Fetched: 2026-03-23 16:54

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.19791v1</id>\n    <title>Text-Based Personas for Simulating User Privacy Decisions</title>\n    <updated>2026-03-20T09:28:15Z</updated>\n    <link href='https://arxiv.org/abs/2603.19791v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.19791v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>The ability to simulate human privacy decisions has significant implications for aligning autonomous agents with individual intent and conducting cost-effective, large-scale privacy-centric user studies. Prior approaches prompt Large Language Models (LLMs) with natural language user statements, data-sharing histories, or demographic attributes to simulate privacy decisions. These approaches, however, fail to balance individual-level accuracy, prompt usability, token efficiency, and population-level representation. We present Narriva, an approach that generates text-based synthetic privacy personas to address these shortcomings. Narriva grounds persona generation in prior user privacy decisions, such as those from large-scale survey datasets, rather than purely relying on demographic stereotypes. It compresses this data into concise, human-readable summaries structured by established privacy theories. Through benchmarking across five diverse datasets, we analyze the characteristics of Narriva's synthetic personas in modeling both individual and population-level privacy preferences. We find that grounding personas in past privacy behaviors achieves up to 88% predictive accuracy (significantly outperforming a non-personalized LLM baseline), and yields an 80-95% reduction in prompt tokens compared to in-context learning with raw examples. Finally, we demonstrate that personas synthesized from a single survey can reproduce the aggregate privacy behaviors and statistical distributions (TVComplement up to 0.85) of entirely different studies.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CR'/>\n    <published>2026-03-20T09:28:15Z</published>\n    <arxiv:primary_category term='cs.CR'/>\n    <author>\n      <name>Kassem Fawaz</name>\n    </author>\n    <author>\n      <name>Ren Yi</name>\n    </author>\n    <author>\n      <name>Octavian Suciu</name>\n    </author>\n    <author>\n      <name>Rishabh Khandelwal</name>\n    </author>\n    <author>\n      <name>Hamza Harkous</name>\n    </author>\n    <author>\n      <name>Nina Taft</name>\n    </author>\n    <author>\n      <name>Marco Gruteser</name>\n    </author>\n  </entry>"
}