Recommendations on the Ethical Use of Novel HIV Data & Analytics

 

As interest in Big Data, machine learning, and AI to improve HIV programs increases, so do critical ethical questions about privacy, surveillance, data misuse, and algorithmic bias. 

To address these concerns and consider frameworks for their responsible use, an 18-member international, interdisciplinary, and intersectoral Working Group, supported by the Bill and Melinda Gates Foundation, convened for multi-day workshops, resulting in the following guidance documents. 

A diverse range of industries and expertise were represented in the working group — including HIV researchers and program innovators with experience in novel data and machine learning; technical experts from computer, information, and data sciences; experts in global public health policy and health law; experts in data ethics and ethics of AI; and members from the community of people living with HIV who have been leaders in advocating for their community. Contributing members are from Malawi, Mozambique, South Africa, the United Kingdom, the United States, and Zimbabwe.

Executive Summary

This guidance pragmatically identifies and addresses key considerations, complexities, and factors of AI and ethics in the context of HIV. Its goals are to:

• help guide researchers and program innovators in building designs that are ethically and technically feasible
• assist those charged with assessing proposals for funding and implementation to effectively assess the proposed use of Big Data and ML 
• identify actions HIV funders and multilateral organizations can take to advance responsible approaches to the use of novel data and ML models in HIV research and programmatic innovation

 

Consensus Factors for Responsible Use of Big Data & AI in HIV Research & Programmatic Innovation

This guidance identifies the key processes, standards, and expectations that researchers, program innovators, and those serving as funding or adoption gatekeepers should address as they navigate ethical considerations in the use of novel data and machine learning methods.

 

Recommendations to Funders & Multi-lateral HIV Organizations

This guidance outlines concrete, actionable recommendations for HIV funders and multilateral organizations to help advance responsible use of novel data and machine learning models in HIV research and programmatic innovation.

 

Full Report

This multi-part guidance provides contextual background on novel data relating to HIV programmatic needs; outlines ethical considerations, concerns, and challenges raised during the working group’s discussion of illustrative scenarios; identifies critical factors for the responsible use of novel data and machine learning in HIV research and programs; and offers concrete recommendations for various stakeholders advancing responsible approaches in this field.