Using Artificial Intelligence to Access Reproductive Healthcare Information by U.S. States Following Sexual Assault

There is currently wide variability across U.S. states regarding the legality of abortion with some states having relatively open access to abortion services and others with a total or near total ban on abortion access. Further complicating these state-by-state differences is the variability in policies and procedures with regard to exemptions (e.g., rape, incest) that may allow women access to abortion services that otherwise would be banned in that state. Factors such as a victim's alcohol use at the time of rape and victim's rape acknowledgment may delay or prevent women from seeking medical care post-assault. To help navigate potential confusion regarding state-level policies and requirements to access abortion services following a rape in which they became pregnant, women may use Artificial Intelligence tools, including ChatGPT, for reproductive health care information. Thus, we sought to qualitatively examine information provided by ChatGPT-4o to women seeking information on how to access abortions in each state depending on their alcohol use and rape acknowledgment after finding out they are pregnant following a rape. Using data collected between October 31 and November 11, 2024, we systematically analyzed ChatGPT-4o's responses to abortion access requests across all 50 U.S. states using four prompts that varied by alcohol use and rape acknowledgment, and qualitatively coded responses for accuracy, helpfulness, and inclusion of empathic statements. Results revealed no significant differences in ChatGPT-4o responses across prompt conditions and one difference based on U.S. state level abortion access restriction. Responses in states with more severe restrictions were more likely to include information on rape exemptions than states with less restrictions in abortion access. However, responses often failed to include helpful information and statements reflective of empathic tone, which may exacerbate pre-existing disparities across U.S. states.