Harvard ReCompute and EAAMO Collaboration

We are proud to share the outstanding work of Harvard ReCompute undergraduate students Audrey Chang, Conan Lu, Gustavo Merino Martinez, Jade Nair, Emmanuel Rassou, and Lia Zheng, who have collaborated with EAAMO in Spring 2024 semester. Their efforts focused on reading and highlighting critical questions from papers presented at previous ACM EAAMO conferences, contributing valuable insights to the broader community. As a result we present to you “Sociotechnical and System Considerations for End-to-end Fairness: An Annotated Reading List.”

In the first week, students explored the use of the AI Incident Database (AIID) to raise awareness of AI harms. Their engagement revealed a significant increase in their understanding of the magnitude of AI harms and the urgency of designing safer AI systems. They also provided actionable recommendations to improve the AIID, emphasizing the need to bridge the gap between technical skills and ethical considerations.

During the second week, the students examined studies addressing bias and fairness in AI systems, discussing feedback loops in automated decision-making and the experiential harms caused by machine learning errors influenced by stereotypes. Their reflections underscored the importance of a comprehensive approach to fairness and bias mitigation in AI. Finally, the students delved into papers on public participation and transparency in AI, analyzing the role of public involvement in democratizing AI and preventing misuse. Their discussions highlighted the need for clear standards, adequate resources, and careful consideration of privacy issues.

Throughout the semester, the Harvard ReCompute undergraduate students demonstrated exceptional dedication and intellectual curiosity. Their work enhances both their own learning and the ongoing discourse on ethical AI practices. We are incredibly proud of their achievements and look forward to their continued impact in the field.

Read the entire annotated reading list here.