TY  - RPRT
TI  - An empirical evaluation of the risks of AI model updates using clinical data: stability, arbitrariness, and fairness
AU  - Ioannis Bilionis
AU  - Ricardo C. Berrios
AU  - Luis Fernandez-Luque
AU  - Carlos Castillo
PY  - 2026
UR  - https://arxiv.org/abs/2604.23954
ID  - 2604.23954
ER  -