Implementation and Adoption of Digital Technologies
Loading...
Paper Type
Complete
Paper Number
2084
Description
Artificial intelligence (AI) based on machine learning technology disrupts how knowledge is gained. Nevertheless, ML’s improved accuracy of prediction comes at the cost of low traceability due to its black-box nature. The field of explainable AI tries to counter this. However, for practical use in IT projects, these two research streams offer only partial advice for AI adoption as the trade-off between accuracy and explainability has not been adequately discussed yet. Thus, we simulate a decision process by implementing three best practice AI-based decision support systems for a high-stake maintenance decision scenario and evaluate the decision and attitude factors using the Analytical Hierarchy Process (AHP) through an expert survey. The combined results indicate that system performance is still the most important factor and that implementation effort and explainability are relatively even factors. Further, we found that systems using similarity-based matching or direct modeling for remaining useful life estimation performed best.
Recommended Citation
Wanner, Jonas; Heinrich, Kai; Janiesch, Christian; and Zschech, Patrick, "How Much AI Do You Require? Decision Factors for Adopting AI Technology" (2020). ICIS 2020 Proceedings. 10.
https://aisel.aisnet.org/icis2020/implement_adopt/implement_adopt/10
How Much AI Do You Require? Decision Factors for Adopting AI Technology
Artificial intelligence (AI) based on machine learning technology disrupts how knowledge is gained. Nevertheless, ML’s improved accuracy of prediction comes at the cost of low traceability due to its black-box nature. The field of explainable AI tries to counter this. However, for practical use in IT projects, these two research streams offer only partial advice for AI adoption as the trade-off between accuracy and explainability has not been adequately discussed yet. Thus, we simulate a decision process by implementing three best practice AI-based decision support systems for a high-stake maintenance decision scenario and evaluate the decision and attitude factors using the Analytical Hierarchy Process (AHP) through an expert survey. The combined results indicate that system performance is still the most important factor and that implementation effort and explainability are relatively even factors. Further, we found that systems using similarity-based matching or direct modeling for remaining useful life estimation performed best.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.