Julie Gerlings is currently doing a PhD at Copenhagen Business School on the topic of automated decisions and xAI (Explainable artificial Intelligence), where se is analyzing the different approaches to xAI to minimize biases in decisions. Besides her PhD, Julie is a consultant at Carve Consulting, where she is putting her knowledge of machine learning and AI into practice for clients.
Biased machines stem from biased data – if we want to build useful machines (ML-models) to support, or make decisions for us, we need to improve the general understanding of why the ML-model operates as it does.
A great deal of work needs to be put into place for us to evaluate machines on even terms with ourselves, as we tend to want them to be perfect. They are not!
We also need to level up our effort to make the automated decisions explainable for the different stakeholders – instead of making workarounds to avoid the machines.
- Automated decision making
- Explainable AI (xAI)
- Human biases and heuristics
In: Handbook of Artificial Intelligence in Healthcare: Vol 2 : Practicalities and Prospects. . ed. /Chee-Peng Lim; Yen-Wei Chen ; Ashlesha Vaidya; Charu Mahorkar ; Lakhmi C. Jain. Cham : Springer 2022, p. 169-198 (Intelligent Systems Reference Library, Vol. 212)
In: Proceedings of the 54th Hawaii International Conference on System SciencesHonolulu : Hawaii International Conference on System Sciences (HICSS) 2021, p. 1284-1293 (Proceedings of the Annual Hawaii International Conference on System Sciences)
In: Effektivitet, No. 2, 2019, p. 37-39
XAI-forsker: Vi ved for lidt om, hvad der egentlig kræves af en god forklaring
Consultant in AI/ML at Carve Consulting