Dies ist eine Übersichtsseite mit Metadaten zu dieser wissenschaftlichen Arbeit. Der vollständige Artikel ist beim Verlag verfügbar.
S951 Revolutionizing Irritable Bowel Disease Management With the Help of Artificial Assistance Technology
0
Zitationen
12
Autoren
2023
Jahr
Abstract
Introduction: Artificial intelligence plays a significant role in the medical field by delivering healthcare information. AI systems can understand and respond to human speech, enabling them to perform various tasks upon command. In this study, 4 popular voice assistants—Apple Siri, Amazon Alexa, Google Assistant, and Microsoft Cortana—were evaluated. Voice recordings were created for 5 frequently asked questions about Ulcerative colitis and Crohn's disease, in 4 languages: English, Spanish, Mandarin, and French. The accuracy of the responses given via audio output and the reliability of the sources supporting each answer was assessed. The researchers determined whether the 4 voice assistants offered clinically sound advice. Methods: This study assessed the reliability of medical information from 4 AI voice assistants—Apple Siri, Amazon Alexa, Google Assistant, and Microsoft Cortana—across English, Spanish, Mandarin, and French for Ulcerative colitis and Crohn's questions. Two blinded authors graded responses in a 2-stage comparison process. The research evaluated response accuracy, supporting source quality, and determined clinically reliable advice while using native speakers and multiple logistic regression analysis. Results: When comparing individual languages, Apple Siri delivered significantly more reliable information in English than Alexa, Google Assistant, and Cortana (98% vs. 88% vs. 45% vs. 13%, P=0.023, 95% CI). Both Alexa and Apple Siri provided dependable information in French compared to other AI systems (88% vs. 78% vs. 45% vs. 10%, P< 0.01, 95% CI). Google Assistant and Apple Siri proved to be more reliable than their counterparts in Spanish (81% vs. 84% vs. 33% vs. 5%, P=0.023, 95% CI) and in Mandarin (91% vs. 88% vs. 21% vs. 1%, P=0.023, 95% CI). AI systems in English yielded significantly better responses compared to those in other languages (88% vs. 57%, P< 0.01) (Figure 1). Conclusion: This study assessed the reliability of 4 AI voice assistants—Apple Siri, Amazon Alexa, Google Assistant, and Microsoft Cortana—in providing healthcare information on Ulcerative colitis and Crohn's disease. Results indicated Siri outperformed others in English, while Siri and Alexa were reliable in French, and Google Assistant and Siri in Spanish and Mandarin. Further research is needed to enhance AI reliability across languages for diverse populations.Figure 1.: Artificial Intelligence Study Outcome.
Ähnliche Arbeiten
Explainable Artificial Intelligence (XAI): Concepts, taxonomies, opportunities and challenges toward responsible AI
2019 · 8.422 Zit.
Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead
2019 · 8.300 Zit.
High-performance medicine: the convergence of human and artificial intelligence
2018 · 7.734 Zit.
Proceedings of the 19th International Joint Conference on Artificial Intelligence
2005 · 5.781 Zit.
Peeking Inside the Black-Box: A Survey on Explainable Artificial Intelligence (XAI)
2018 · 5.519 Zit.