Essays on artificial intelligence (AI) in management

• Main Author: ZHOU, Bowen
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2024
• Subjects: Aritificial intelligence; Performance; Geopolitical conflicts; Solution finding; Artificial Intelligence and Robotics; Strategic Management Policy; Technology and Innovation
• Online Access: https://ink.library.smu.edu.sg/etd_coll/631; https://ink.library.smu.edu.sg/context/etd_coll/article/1629/viewcontent/ESSAYS_ON_ARTIFICIAL_INTELLIGENCE_1_.pdf
• Institution: Singapore Management University

This dissertation comprises three essays that investigate the transformative potential of Artificial Intelligence (AI) in business. Chapter 1 investigates the fundamental issue of how integrating AI within R&D activities influences a firm’s market value. We developed an "AI Index" using patent data and textual analysis. Interestingly, empirical results indicate a negative correlation between AI integration and market value. However, this does not suggest that AI is an unviable avenue for exploration. Further analysis of the boundary conditions reveals that complementary assets are crucial for successful commercialisation, highlighting that while AI adoption is costly, these assets significantly enhance its market value. In Chapter 2, my research has examined how firms have adapted their R&D activities to incorporate AI as a strategy to mitigate potential adversities arising from such conflicts. We leveraged the 2018 bans on Huawei as an exogenous shock and utilised a difference-in- differences model to evaluate the effects. The findings indicate that geopolitical conflicts have a positive impact on the adoption of AI in firms' R&D activities, as it enables them to preemptively address potential future restrictions. Chapter 3 focuses on a more micro-level analysis, specifically on developers, examining how the advent of ChatGPT affects knowledge searching. Using Stack Overflow as a context, which separates question formulation from problem-solving, we conducted an exploratory-style empirical test. This study reveals that while AI-generated content technologies like ChatGPT provide more potential solutions, these do not necessarily translate into accepted solutions. Additionally, we discovered that the presence of AI increases the time required to evaluate these solutions. We also considered varying capabilities by examining search depth and scope, finding that AI benefits non-domain experts by reducing the learning curve costs.