Artificial Intelligence and Women’s Right to Health in Crisis and Post-Conflict Settings: Human Rights Requirements for Data-Driven Decision-Making
Published in National Conference on Artificial Intelligence, Human Rights and Peace (in Persian), 2026
Nazanin Azizmohammadi, Ph.D.
Senior Research Affiliate, Dataism Lab, Virginia Tech
Conference paper: National Conference on Artificial Intelligence, Human Rights and Peace.
Language: Full text in Persian; English abstract included. The PDF contains the full article, assessment framework and references (17 pages).
Download the full article (PDF, in Persian)
Abstract
Artificial intelligence can accelerate the review of records, clinical prioritisation and allocation of health services during crisis and post-conflict recovery. Yet records produced in these settings reflect disrupted documentation and prior inequalities in access. Women affected by displacement, disability, insecurity or caregiving responsibilities may be underrepresented, allowing an absence of information to be mistaken for an absence of need. This article asks what legal requirements should govern data-driven decisions to prevent this gap from becoming exclusion from essential care. Using doctrinal and normative analysis, it examines Iranian law, the International Covenant on Economic, Social and Cultural Rights, the Convention on the Rights of Persons with Disabilities and humanitarian law. Three hypothetical scenarios concerning pregnancy care, mental health after violence and treatment allocation show why aggregate accuracy, removal of the gender variable and nominal human oversight cannot alone guarantee equality. The article proposes a nine-stage gender-sensitive right-to-health impact assessment covering necessity and consequences, data coverage and absence, definitions and proxies for need, disaggregated performance, privacy, participation, human oversight, review and remedy, and monitoring and exit. Its contribution connects data absence with indirect discrimination, evaluates decisions against the availability, accessibility, acceptability and quality of care, and requires renewed assessment after emergency use. Evidence concerning language-model summaries reinforces the need to test each tool in its intended setting. Emergency conditions do not remove core health obligations or the duty of non-discrimination.
Keywords: Artificial intelligence; Women’s right to health; Gender equality; Data governance; Crisis and post-conflict settings