Consent or Compromise? The Hidden Costs of AI in Cervical Cancer Screening

Tanzania is one of nine global country sites collecting data from cervical cancer screening that is being used to further develop, train and validate a deep learning algorithm to be deployed as software on devices that take cervical images. In an era when AI holds a seemingly endless horizon of potential to revitalize underserved healthcare systems, is the race to develop it leading us to deprioritize the security of those systems and the rights of the patients within them?

Karen Yeates, one of the authors of this case study and research scientist studying strategies to improve cervical cancer prevention in East Africa, has been leading a team of researchers in Tanzania that has been implementing several research projects within the Human Papillomavirus Automated Visual Evaluation (PAVE) project.

The case presents a multi-faceted challenge of balancing technological innovation with ethical responsibility. Participants are encouraged to evaluate the ethical, technical and operational dimensions of deploying AI in global health initiatives.