Influencing Factors on the Adoption of AI: Insights From Social, Organizational, Individual and Methodological Perspectives

Artificial Intelligence (AI) applications are becoming increasingly important, simplifying daily life and supporting organizations across various applications. Despite the numerous positive attributes and the potential of AI, studies indicate that adoption rates, particularly within organizations, are still not as high as expected. To increase the utilization of AI, factors such as the user’s culture, industry-specific elements, willingness to pay, and psychological factors play a critical role. By leveraging factors that promote AI adoption and addressing barriers, it is possible to enhance the integration of AI technologies. This dissertation examines the factors influencing AI adoption from four perspectives: (1) social, (2) organizational, (3) individual, and (4) methodological. From the social perspective, one study included in this dissertation investigates the influence of culture on the adoption of AI as an emerging technology among others. A structured literature review (SLR) was conducted, focusing on Information Systems (IS) papers from the Basket of Eight that measure the effect of culture. The knowledge extracted from these papers was then condensed, with existing research categorized by research areas, data collection methods, and their assessments. The resulting concept matrix serves as a valuable summary and foundation for future research. A recommendation from the SLR is that future research should measure culture individually rather than making assumptions based on the country a person lives in. Additionally, the study explicitly provides a research agenda highlighting existing gaps in the literature. For instance, the cultural impact on the adoption of newer forms of AI, such as Generative AI (GenAI), should be measured. Thus, this study demonstrates the significant influence of culture as a social factor on AI adoption. For the organizational perspective, two studies were conducted. The first study utilizes a case study approach, consisting of seven interviews within the financial services and manufacturing industries. From these interviews, drivers and barriers of AI adoption were identified and categorized using the Technology-Organization-Environment (TOE) framework. The factors were also classified by industry and compared between the two industries. This leads to the identification of soft factors that are industry-specific and hard factors that are more general. For example, a soft barrier specific to the financial industry is the presence of legacy systems, while a general driver is the potential for cost reduction. The identified factors are particularly useful for organizations within these industries, but the more general factors can be applied to other organizations as well. The second study within the organizational perspective examined the willingness to pay (WTP) for machine learning-based software testing tools in organizations using a conjoint analysis. Initially, attributes important for these tools and their target audience were identified through a structured literature review and a Delphi study. Attributes such as accuracy, ease of use, and integration were found to be crucial for the adoption. The conjoint analysis, conducted with 119 software testers in Germany, revealed that they are willing to pay up to €120 more per license per month for an increase in accuracy from 90% to 99%. This study highlights WTP as an adoption factor when introducing AI and identifies the essential attributes AI systems must possess to be successfully adopted. Thus, the organizational perspective uncovers various influencing factors examined within this dissertation. Also, two studies were conducted from the individual perspective. Both utilized online experiments to investigate the psychological factors influencing AI adoption. The first study in this perspective examined the impact of ChatGPT assistance on the performance and perceived meaningfulness among programmers. The study involved 161 experienced coders who completed coding and debugging tasks with and without ChatGPT assistance. The results showed a significant increase in performance but a decrease in perceived meaningfulness due to the reduced difficulty of the tasks. As the results are depending on the tasks, the adoption of AI should be carefully considered as lower meaningfulness can result in less motivation for work. Another study in this perspective, involving 174 participants from Germany, demonstrated that the IKEA effect occurs when using Generative AI. This effect arises when individuals value an output more if they invested more effort in its creation. This suggests that instead of solely aiming for automation, which is often seen as the goal of AI, fostering collaboration could enhance AI adoption. These studies collectively demonstrate the significant influence of individual psychological factors on AI adoption. Finally, the methodological perspective emphasizes the importance of thorough and sound methodology. Using a Design Science Research approach, this study initially examines how previous research has conducted and reported cross-sectional surveys. Based on this analysis, focus groups were utilized to gain deeper insights. This results in eleven guidelines which serve as a foundation to allow particularly inexperienced researchers to conduct their (AI adoption) research in a reproducible and reliable manner. Overall, these studies show that various factors influence the adoption of AI, which must be considered from different perspectives. Only by examining these different angles AI can be successfully implemented. Thus, AI does not only simplify daily life but also supports organizations in improving or streamlining processes. Moreover, the role of society is also crucial, as well as a solid methodology to make research generalizable.