A Question of Trust: Old and New Metrics for the Reliable Assessment of Trustworthy AI

Andrea Campagner, Riccardo Angius, Federico Cabitza, Federico Cabitza

2023

Abstract

This work contributes to the evaluation of the quality of decision support systems constructed with Machine Learning (ML) techniques in Medical Artificial Intelligence (MAI). In particular, we propose and discuss metrics that complement and go beyond traditional assessment practices based on the evaluation of accuracy, by focusing on two different dimensions related to the trustworthiness of a MAI system: reputation/ability, which relates to the accuracy or predictive ability of the system itself; and expertise/source reliability, which relates instead to the trustworthiness of the data which have been used to construct the MAI system. Then, we will discuss some previous, but so far mostly neglected, proposals as well novel metrics, visualizations and procedures for the sound evaluation of a MAI system’s trustworthiness, by focusing on six different concepts: advice accuracy, advice reliability, pragmatic utility, advice value, decision benefit and potential robustness. Finally, we will illustrate the application of the proposed concepts through two realistic medical case studies.

Paper Citation

in Harvard Style

Campagner A., Angius R. and Cabitza F. (2023). A Question of Trust: Old and New Metrics for the Reliable Assessment of Trustworthy AI. In Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 5: HEALTHINF; ISBN 978-989-758-631-6, SciTePress, pages 132-143. DOI: 10.5220/0011679600003414

in Bibtex Style

@conference{healthinf23,
author={Andrea Campagner and Riccardo Angius and Federico Cabitza},
title={A Question of Trust: Old and New Metrics for the Reliable Assessment of Trustworthy AI},
booktitle={Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 5: HEALTHINF},
year={2023},
pages={132-143},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011679600003414},
isbn={978-989-758-631-6},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 16th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2023) - Volume 5: HEALTHINF
TI - A Question of Trust: Old and New Metrics for the Reliable Assessment of Trustworthy AI
SN - 978-989-758-631-6
AU - Campagner A.
AU - Angius R.
AU - Cabitza F.
PY - 2023
SP - 132
EP - 143
DO - 10.5220/0011679600003414
PB - SciTePress