AI Project failure: a conceptual framework of organizational challenges

Abstract

This paper provides a detailed and comprehensive examination of the organizational challenges that lead to the typical incidence of AI project failure. A systematic literature review (SLR) was conducted. Using the building block and snowballing approach, an SLR was conducted. As a result, a theoretical framework with thirteen organizational challenges has been developed. An empirical study was conducted to validate the framework, including interviews with subject-matter experts. The findings from the interviews supported the veracity of the framework's identified issues while revealing new data-related difficulties. Each obstacle has been identified as significant or consequential in practical use. Early in a project's lifespan, problems with regulating expectations, such as dreaming too big, a lack of knowledge, a lack of budget, a lack of understanding of business and user objectives, and a misunderstanding of AI's capabilities, are common. Problems with cross-organizational collaborations, such as those requiring change management, Lack of collaboration, a lack of support from top management, and cultural opposition, frequently emerge during the intermediate periods of a project's growth. Data issues, such as data quality, access, and security, come later in the AI development process. This study of the organizational elements contributing to artificial intelligence initiatives' failure yielded ethical and credible results.

Date of Award	4 Mar 2023
Original language	English
Supervisor	Samaneh Bagheri (Examiner) & Khoi Nguyen (Co-assessor)