From Turing Test to Chinese Room Argument: how to apply artificial intelligence in aviation

Simple item page
dc.contributor.authorSaunders, Declan
dc.contributor.authorLi, Wen-Chin
dc.contributor.authorWang, Thomas
dc.date.accessioned2025-07-03T20:48:07Z
dc.date.available2025-07-03T20:48:07Z
dc.date.freetoread2025-07-03
dc.date.issued2025
dc.date.pubOnline2025-05-20
dc.descriptionEuropean Association for Aviation Psychology Conference EAAP 35
dc.description.abstractThe emergence of artificial intelligence (AI) with advanced large language model (LLM) offers promising approaches for enhancing the capacity of textual analysis. Both the Turing Test and Chinese Room Argument explore AI’s understanding of human language, although both methodologies have dissimilar interpretation on AI’s ‘intelligence’. Current AI systems have demonstrated the capacity for achieving defined test goals for ‘intelligence’. The aviation industry is increasingly interested in adopting AI to improve efficiency, safety, and cost efficiency; with the Generative Pre-trained Transformers’ (GPT) capability to reduce resource-intensive analytics in accident causation classification. This study investigates the potential and challenges of using AI to analyze human factors involved in aviation accidents based on the Human Factors Analysis and Classification System (HFACS). Six subject-matter experts in aviation human factors and AI domain participated in this research. All participants were familiar with the HFACS framework to analyze aviation accident reports and the output of GPT which were based on the prompt engineering developed by the research team. This research creates a framework to perform its initial generation and training using GE 235 accident investigation report from Taiwan Transportation Safety Board (TTSB). Initial discoveries demonstrated that the AI model could populate the sub-dimensions of Level 1 HFACS framework with moderate accuracy, although there remains a high presence of hallucinations in generated outputs, with a lack of reproducibility in consecutive outputs with consistent input data. There are still different opinions on AI applications in real-world operations with ethics and safety concerns. While there is clear potential for GPT models to supplement accident analysis within the HFACS framework, there is still more work to integrate HFACS framework into GPT modelling for effective generation to accident data.
dc.description.journalNameTransportation Research Procedia
dc.format.extentpp. 270-277
dc.identifier.citationSaunders D, Li W-C, Wang T. (2025) From Turing Test to Chinese Room Argument: how to apply artificial intelligence in aviation. Transportation Research Procedia, Volume 88, 2025, pp. 270-277en_UK
dc.identifier.elementsID673331
dc.identifier.issn2352-1465
dc.identifier.urihttps://doi.org/10.1016/j.trpro.2025.05.033
dc.identifier.urihttps://dspace.lib.cranfield.ac.uk/handle/1826/24154
dc.identifier.volumeNo88
dc.languageEnglish
dc.language.isoen
dc.publisherElsevieren_UK
dc.publisher.urihttps://www.sciencedirect.com/science/article/pii/S2352146525004120?via%3Dihub
dc.rightsAttribution 4.0 Internationalen
dc.rights.urihttp://creativecommons.org/licenses/by/4.0/
dc.subject3505 Human Resources and Industrial Relationsen_UK
dc.subject35 Commerce, Management, Tourism and Servicesen_UK
dc.subjectNetworking and Information Technology R&D (NITRD)en_UK
dc.subjectMachine Learning and Artificial Intelligenceen_UK
dc.subject3509 Transportation, logistics and supply chainsen_UK
dc.subjectAccident Analysisen_UK
dc.subjectArtifical Intelligenceen_UK
dc.subjectChinese Room Argumenten_UK
dc.subjectHuman Factors Analysisen_UK
dc.subjectClassification Systemen_UK
dc.subjectLarge Language Modelen_UK
dc.titleFrom Turing Test to Chinese Room Argument: how to apply artificial intelligence in aviationen_UK
dc.typeArticle

Files

Original bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
From_Turing_Test-2025.pdf
Size:
438.74 KB
Format:
Adobe Portable Document Format
Description:
Published version
Download
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.63 KB
Format:
Plain Text
Description:
Download

Collections

Staff publications (AA)