… aspect of AI. In this paper, we report on data practices in high-stakes AI, from interviews
with 53 AI … We defne, identify, and present empirical evidence on Data Cascades—compounding …

… on data practices in high-stakes AI, from interviews with 53 AI … desiderata within high-stakes
domains—data reliability and … , not the data work": Data cascades in high-stakes AI. Proc. of …

… in part to allow for their use in AI (artificial intelligence) applications. Just as those involved in
… with AI, the developers of AI systems should pay attention to issues that affect the data upon …

… [39] state that “Data quality carries an elevated significance in high-stakes AI due to its …
“Everyone wants to do the model work, not the data work”: Data Cascades in High-Stakes AI…

… framework describing how data scientists develop mental models of AI behavior. By framing
AI analysis as sensemaking, we aim at providing a language for describing AI analysis, help …

… Data work. Our results indicate the sobering prevalence of messy, protracted, and opaque
data cascades in high-stakes domains. Data cascades point to the contours of a larger …

… how, in high-stakes domains like autonomous vehicles, data … Feedback delays contribute to
“data cascades,” or … frankly I think that’s how people should be doing AI. It [shouldn’t be] like: …