Post by Pablo C.

Great to be back from EURO Leeds, where I presented on fairness stability and its connection to Explainable AI. The discussions with fellow operational researchers were incredibly insightful! It got me thinking: when engaging with regulators, how can we better highlight the limitations of metrics rather than just their intuitive logic? For instance, some fairness metrics I introduced to the credit industry are now seeing some modest adoption, which is fantastic! Yet, it's imperative we ensure that everyone – especially those outside OR – recognizes that all metrics have inherent shortcomings and aren't one-size-fits-all solutions. Understanding these nuances is key to preventing unintended consequences as AI is taking a more active and central role in today’s world. If you want to discuss anything regarding my research and work please don’t hesitate to reach out to me. I would like to thank my supervisory team (Yu Huan and Christophe Mues) for guiding me through this process. I would also like to thank Stefan Lessmann Sven Weinzierl and Nico Hambauer for the wonderful debates on the future of OR AI and (surprisingly) applied physics. #OR #AIethics #FairnessMetrics #AIAct #EUROLeeds #CreditScoring #ModelRisk #EURO2025 #OperationsResearch