Red Onions, Soft Cheese and Data: From Food Safety to Data Traceability for Responsible AI

Abstract

Software systems that learn from data with AI and machine learning (ML) are becoming ubiquitous and are increasingly used to automate impactful decisions. The risks arising from this widespread use of AI/ML are garnering attention from policy makers, scientists, and the media, and lead to the question what data management research can contribute to reduce such risks. These dangers of AI/ML applications are relatively new and recent, however our societies have had to deal with the dangers of complex and distributed technical processes for a long time already. Based on this insight, we detail how the U.S. Food and Drug Administration (FDA) combats the outbreaks of foodborne illnesses, and use their processes as an inspiration for a data-centric vision towards responsible AI.