Trustworthy artificial intelligence: A decision‐making taxonomy of potential challenges
Open Access
- 31 May 2023
- journal article
- research article
- Published by Wiley in Software: Practice and Experience
- Vol. 54 (9) , 1621-1650
Abstract
The significance of artificial intelligence (AI) trustworthiness lies in its potential impacts on society. AI revolutionizes various industries and improves social life, but it also brings ethical harm. However, the challenging factors of AI trustworthiness are still being debated. This research explores the challenging factors and their priorities to be considered in the software process improvement (SPI) manifesto for developing a trustworthy AI system. The multivocal literature review (MLR) and questionnaire‐based survey approaches are used to identify the challenging factors from state‐of‐the‐art literature and industry. Prioritization based taxonomy of the challenges is developed, which reveals that lack of responsible and accountable ethical AI leaders, lack of ethics audits, moral deskilling & debility, lack of inclusivity in AI multistakeholder governance, and lack of scale training programs to sensitize the workforce on ethical issues are the top‐ranked challenging factors to be considered in SPI manifesto. This study's findings suggest revising AI‐based development techniques and strategies, particularly focusing on trustworthiness. In addition, the results of this study encourage further research to support the development and quality assessment of ethics‐aware AI systems.Keywords
This publication has 57 references indexed in Scilit:
- Social research 2.0: virtual snowball sampling method using FacebookInternet Research, 2012
- The Emergence and Maintenance of Ethical Leadership in OrganizationsJournal of Personnel Psychology, 2012
- Identifying relevant studies in software engineeringInformation and Software Technology, 2011
- Factors influencing clients in the selection of offshore software outsourcing vendors: An exploratory study using a systematic literature reviewJournal of Systems and Software, 2011
- A systematic review of search-based testing for non-functional system propertiesInformation and Software Technology, 2009
- Guidelines for conducting and reporting case study research in software engineeringEmpirical Software Engineering, 2008
- Using analytic hierarchy process (AHP) to improve human performance: An application of multiple criteria decision making problemJournal of Intelligent Manufacturing, 2004
- Construction Partnering Process and Associated Critical Success Factors: Quantitative InvestigationJournal of Management in Engineering, 2002
- Principles of survey researchACM SIGSOFT Software Engineering Notes, 2002
- A statistical approach to designing search filters to find systematic reviews: objectivity enhances accuracyJournal of Information Science, 2001