Pular para o conteúdo principal

ETHNOS_APP

Início • Busca • Periódicos • Lista 0

In Defense of Post Hoc Explanations in Medical AI

Dados Bibliográficos

ID17536173
AutoresJoshua Hatherley (0000-0002-8581-9669), Lauritz Aastrup Munch (0000-0002-3510-5422), Jens Christian Bjerring (0000-0001-8755-6746)
Ano2026
Volume56
Fascículo1
Páginas40-46
Data de publicação2026-01-01
Peer ReviewedSim
Open AccessSim
TipoARTICLE
PeriódicoThe Hastings Center Report (JOURNAL)
Identificadores do periódicoISSN: 0093-0334 • E-ISSN: 1552-146X
EditoraWiley (PUBLISHER • GB)
DOI10.1002/hast.4971
PMID41639026
OpenAlexW7128078825
IdiomaEN
Referências citadas42

Since the early days of the explainable artificial intelligence movement, post hoc explanations have been praised for their potential to improve user understanding, promote trust, and reduce patient-safety risks in black box medical AI systems. Recently, however, critics have argued that the benefits of post hoc explanations are greatly exaggerated since they merely approximate, rather than replicate, the actual reasoning processes that black box systems take to arrive at their outputs. In this paper, we aim to defend the value of post hoc explanations against this recent critique. We argue that even if post hoc explanations do not replicate the exact reasoning processes of black box systems, they can still improve users' functional understanding of black box systems, increase the accuracy of clinician-AI teams, and assist clinicians in justifying their AI-informed decisions. While post hoc explanations are not a silver-bullet solution to the black box problem in medical AI, they remain a useful strategy for addressing it

Black box · Post hoc · Post-hoc analysis · Replicate · Wireless ad hoc network · Artificial Intelligence in Healthcare and Education · Clinical Reasoning and Diagnostic Skills · Explainable Artificial Intelligence (XAI

  • To Trust or to Think

    Open Access•Zana Buçinca, Maja Barbara Malaya et al.•Proceedings of the ACM on…•2021

  • Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead

    Open Access•Cynthia Rudin•Nature Machine Intelligence•2019

  • On the ethics of algorithmic decision-making in healthcare

    Open Access•Thomas Grote, Philipp Berens•Journal of Medical Ethics•2020

  • A Survey of Methods for Explaining Black Box Models

    Open Access•Riccardo Guidotti, Anna Monreale et al.•ACM Computing Surveys•2019

  • High-performance medicine

    Open Access•Eric J Topol•Nature Medicine•2019

  • Putting explainable AI in context

    Open Access•Mark Theunissen, Jacob Browning•Ethics and Information Technology•2022

  • AI and the need for justification (to the patient)

    Open Access•Anantharaman Muralidharan, Julian Savulescu et al.•Ethics and Information Technology•2024

  • Conceptualizing understanding in explainable artificial intelligence (XAI)

    Open Access•Timo Speith, Barnaby Crook et al.•Ethics and Information Technology•2024

  • The Virtues of Interpretable Medical AI

    Open Access•Joshua Hatherley, Robert Sparrow et al.•Cambridge Quarterly of Healthcare…•2024

  • Defining the undefinable

    Jordan Joseph Wadden•Journal of Medical Ethics•2022

  • On the Opacity of Deep Neural Networks

    Open Access•Anders Søgaard•Canadian Journal of Philosophy•2023

  • Against Interpretability

    Open Access•Maya Krishnan•Philosophy & Technology•2020

  • Artificial Intelligence and Patient-Centered Decision-Making

    Open Access•Jens Christian Bjerring, Jacob Busch•Philosophy & Technology•2021

  • Transparency in Algorithmic and Human Decision-Making

    Open Access•John Zerilli, Alistair Knott et al.•Philosophy & Technology•2019

  • The mindlessness of ostensibly thoughtful action

    Ellen J Langer, Arthur Blank et al.•Journal of Personality and Social…•1978

Velocidade de citaçãohistorical
Altamente citadoNão
Ethnos_APP • Projeto Open Source • Licença MIT • Frontend v2.0.0 • Privacidade e Cookies • Documentação da API: api.ethnos.app/docs • Código da API: GitHub • DOI: 10.5281/zenodo.17049435 • Código do Frontend: GitHub • DOI: 10.5281/zenodo.17050053 • cruz.rio.br • Expectantes Misericordiae