Getting ahead of bias in AI applications

Catalog description

Artificial Intelligence (AI), the science and engineering of achieving tasks historically associated with human intelligence, is everywhere. As it rapidly evolves, staying ahead of ethical challenges is a new frontier. Society will need philosophers, ethicists, and creatives to have or partner with computer science and engineering skillsets to face the new realities that emerge as machines can learn from data to guide and make decisions. Nowhere is this more true than in the delivery of healthcare. In healthcare, AI is already filling roles like helping clinicians better identify frail patients at risk of poor health outcomes. But how can AI be a better part of the team? And whose responsibility is it when AI gets it wrong? In this course, we will take examples across the human life span and explore how AI can enhance human quality of life and experience. We will also explore where disparities in access to AI hinder outcomes and where bias in AI algorithms can lead to inappropriate differential treatment of patients. We will use qualitative interview methods as prework to learn approaches for getting ahead of bias in the development of AI. We will also explore principles of co-design, and consider how to bring the perspectives of patients and caregivers into the development of future AI approaches.