Privacy and Safety Experiences and Concerns of US Women Using Generative AI for Seeking Sexual and Reproductive Health Information

Paper Title

Privacy and Safety Experiences and Concerns of US Women Using Generative AI for Seeking Sexual and Reproductive Health Information

Publication Info

• Topic area: Privacy and safety in the use of generative AI for sexual and reproductive health information.
• Keywords: Generative AI, sexual and reproductive health, privacy, safety, user experiences, abortion, data practices, surveillance, mitigation strategies, policy recommendations.

Background and Problem

• Problem / challenge: Existing research on generative AI (GenAI) chatbots focuses on technical and clinical aspects, neglecting user privacy and safety concerns, especially in sensitive contexts like sexual and reproductive health (SRH).
• Significance: With the overturning of Roe v. Wade, privacy risks in seeking SRH information have increased, particularly for abortion-related queries, as digital data could lead to criminalization or surveillance.
• Motivation and related work: Prior studies have explored privacy in FemTech tools like period-tracking apps but have not addressed GenAI chatbots. The conversational nature of GenAI tools introduces unique risks, such as excessive data collection, profiling, and government access, which remain underexplored.

Solution

• Proposed approach: Semi-structured interviews with 18 U.S.-based participants were conducted to understand privacy and safety experiences when using GenAI chatbots for SRH information.
• Novelty:
  1. First study to examine privacy and safety concerns of GenAI chatbot users in the SRH context.
  2. Identification of unique risks, such as government surveillance and profiling, specific to GenAI chatbots.
  3. Recommendations for socio-technical and policy interventions to enhance privacy and safety.
• Procedure and key techniques:
  • Participants were recruited from both restrictive and non-restrictive states.
  • Interviews explored factors influencing adoption, beliefs about data practices, privacy risks, and mitigation strategies.
  • Thematic analysis was used to identify key themes and sub-themes.

Results

• Concrete findings:
  • Participants valued GenAI chatbots for their utility, usability, accessibility, and perceived empathy but noted limitations in handling serious health issues and providing personalized advice.
  • Privacy concerns included excessive data collection, government surveillance, profiling, data use for model training, and lack of regulatory protections.
  • Abortion-related queries elicited heightened concerns, particularly in restrictive states.
• Advantage over baselines: The study highlights privacy risks specific to GenAI chatbots, which are more interactive and anthropomorphic than other SRH tools like search engines or period-tracking apps.
• Experiments / evaluation:
  • 18 semi-structured interviews with participants aged 18–45 who had used GenAI chatbots for SRH information.
  • Participants were evenly split between restrictive and non-restrictive states.
  • Data were analyzed using thematic analysis to identify patterns in adoption, privacy concerns, and mitigation strategies.
• Limitations and future work:
  • Limited to U.S.-based participants aged 18–45; findings may not generalize to older adults or non-U.S. contexts.
  • Small sample size limits statistical representativeness.
  • Future work should include quantitative studies and explore non-users’ perspectives.

Summary

This study investigates the privacy and safety experiences of U.S. women using GenAI chatbots for SRH information, particularly in the post-Roe v. Wade era. Participants valued the tools for their accessibility and perceived empathy but expressed concerns about excessive data collection, profiling, and government surveillance, especially for abortion-related queries. Few participants employed effective privacy protections, highlighting the need for better user education and system design. The findings inform recommendations for user-centered design, privacy-by-design principles, interactive privacy features, and stronger regulatory protections to ensure safe and ethical use of GenAI in sensitive contexts.