This article explores Human-Centered Artificial Intelligence (HCAI) in medical cytology, with a focus on enhancing the interaction with AI. It presents a Human–AI interaction paradigm that emphasizes explainability and user control of AI systems. It is an iterative negotiation process based on three interaction strategies aimed to (i) elaborate the system outcomes through iterative steps (Iterative Exploration), (ii) explain the AI system’s behavior or decisions (Clarification), and (iii) allow non-expert users to trigger simple retraining of the AI model (Reconfiguration). This interaction paradigm is exploited in the redesign of an existing AI-based tool for microscopic analysis of the nasal mucosa. The resulting tool is tested with rhinocytologists. The article discusses the analysis of the results of the conducted evaluation and outlines lessons learned that are relevant for AI in medicine.

A Human–AI interaction paradigm and its application to rhinocytology

Giuseppe Desolda
;
Giovanni Dimauro;Andrea Esposito;Rosa Lanzilotti;
2024-01-01

Abstract

This article explores Human-Centered Artificial Intelligence (HCAI) in medical cytology, with a focus on enhancing the interaction with AI. It presents a Human–AI interaction paradigm that emphasizes explainability and user control of AI systems. It is an iterative negotiation process based on three interaction strategies aimed to (i) elaborate the system outcomes through iterative steps (Iterative Exploration), (ii) explain the AI system’s behavior or decisions (Clarification), and (iii) allow non-expert users to trigger simple retraining of the AI model (Reconfiguration). This interaction paradigm is exploited in the redesign of an existing AI-based tool for microscopic analysis of the nasal mucosa. The resulting tool is tested with rhinocytologists. The article discusses the analysis of the results of the conducted evaluation and outlines lessons learned that are relevant for AI in medicine.
File in questo prodotto:
File Dimensione Formato  
1-s2.0-S0933365724001751-main.pdf

accesso aperto

Tipologia: Documento in Versione Editoriale
Licenza: Creative commons
Dimensione 3.11 MB
Formato Adobe PDF
3.11 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11586/498000
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 1
  • ???jsp.display-item.citation.isi??? ND
social impact