Lorna Downie, a PhD Candidate at the KIN Center for Digital Innovation, received the Best Student Paper Award at the Academy of Management's (AOM) 2026 annual meeting.
The paper examines how organizations manage predictive models over time, specifically when those models are used for high-stakes decisions. Predictive models are tools, often algorithmic or data-driven, that forecast future outcomes based on patterns in past data, for example, estimating which employee is likely to perform best, or which athlete is most likely to succeed at a major competition. While companies increasingly rely on tools like these to guide major choices, little is known about how organizations handle it when a model's predictions drift out of step with reality. Downie's paper addresses that gap directly.
Drawing on a four-year ethnographic study of a national Olympic committee and three national sport federations, the paper analyzes how predictive models are used to inform Olympic athlete selection.
Downie (AI@Work research group) and her co-authors introduce the concept of "ground truthing," a cyclical, sociotechnical practice through which organizations assess and revise their predictive models. This cycle unfolds in three phases: reflecting, where organizations look back at how a model's predictions held up; resetting, where the model is recalibrated based on what was learned; and running, where the revised model is put back into use, until the next opportunity to check it against reality. In this case, that opportunity is anchored by the Olympic Games themselves, a fixed and universally recognized moment where predictions are tested against what actually happens.
The paper also introduces a second concept, "model stewarding," to capture a broader kind of organizational work: not just checking whether a model is statistically accurate, but interrogating the values built into it, examining its real-world consequences for the people affected by it, and revising it to better serve the outcomes an organization actually wants. Together, the two concepts show how organizations can stabilize predictive models for high-stakes decisions while remaining attentive to the values and broader effects those models have over time.
This marks Downie's second Best Student Paper Award from AOM. In 2023, she won for her paper 'On the right track? Studying the use of biometric data to manage people in a sports organization,' which drew on a 16-month ethnographic investigation into the use of biometric data in an elite sport setting. That study revealed a new form of control enacted through biometric data, what the authors described as a shared control of the body.