Erin Kilbride (she/they) is a Researcher at Human Rights Watch, investigating violence against sex workers, LGBT rights, and human trafficking. She is also an Adjunct Professor in the Women’s & Gender Studies Program at Georgetown University, where she was previously a Visiting Researcher. Erin has authored dozens of global investigations into military trials of civilians, targeted killings of human rights defenders, migrant rights, and sexual violence perpetrated by security forces. Prior to joining Human Rights Watch, Erin was the Research Coordinator at Front Line Defenders, providing emergency support to human rights defenders at risk and leading fact-finding missions across Africa, Asia, the Americas, and MENA regions. Erin has worked on border demilitarization and Search & Rescue missions with several migrant solidarity collectives, and her reporting has appeared in The Guardian, Al Jazeera, NPR, HuffPost, Wall Street Journal, and The Hill.
Research Project:”Tech Against Trafficking”: Risks & Opportunities of AI-Powered Anti-Trafficking Operations
This research explores how and to what consequence the use of Artificial Intelligence (AI) to fight human trafficking will reproduce the methodological flaws, inefficacies, and violence endemic to past waves of multinational anti-trafficking operations. It situates AI tools built to fight trafficking inside longstanding feminist critiques of “Rescue and Raid,” and maps the adoption of such tools by US government agencies. For decades, anti-trafficking operations have relied on racial, gender, and class stereotypes as proxy indicators of exploitation. If AI systems are built using the same tropes, they risk furthering disproportionate surveillance of marginalized people and undermining global efforts to eradicate exploitation. In its 2024 Trafficking in Persons Report, the State Department called for the development of “data and algorithm tools to detect human trafficking patterns, identify suspicious and illicit activity, and report” these to law enforcement. The report said AI language models should be used to “detect, translate, and categorize key words used by traffickers.” A 2022 literature review of computing research, however, found a “general lack of ground truth data sets especially with respect to distinguishing between consensual sex work and human trafficking,” which prompts developers to “rely on proxy indicators of trafficking such as suspicious key words” generated by law enforcement (Deeb-Swihart et al.). Studies overwhelmingly show that laws, policies, and digital tools which conflate sex work with trafficking are counterproductive and ineffective. This project builds on several years of field research analyzing how criminalization and over-policing of sex workers undermine anti-trafficking efforts. The methodology centers sex workers as experts in identifying exploitation, based on interviews conducted with more than 600 sex workers and survivors of trafficking in dozens of countries.