Do AI-Based Tools Increase Task Performance: The Mediating Role of Trust and Discomfort


Thesis Type: Postgraduate

Institution Of The Thesis: Marmara University, Institute of Social Sciences, Department of Business Administration (Eng), Turkey

Approval Date: 2024

Thesis Language: English

Student: ABDULKADİR FURKAN AYHAN

Supervisor: Hüseyin Ekizler

Abstract:

In the 1940s, McCulloch & Pitts laid the foundations of artificial intelligence with their model of the brain as a Boolean circuit. Views on artificial intelligence, which are said to have been established on this basis, are evolving, and changing every day. First coined in 1956, Artificial Intelligence has become a subject of conferences over the years, evolving into an industry and a science. Especially after AlphaGo, artificial intelligence applications that attract more attention from users have become a topic of curiosity for many people today, including employers, employees, and others (Russell, Norvig, 1995).

With the rapid development and change in technology, organizations are increasingly focusing on the widespread digitization. One of the most important topics within digitization today is applications based on artificial intelligence (AI). When looking at the current graphs shared within the study, it can be seen that there is an increasing trend towards AI topics.

In this context, organizations are entering into optimization processes in their current processes through AI-based applications, conducting various studies to reduce their expenses within the scope of cost optimization. While AI applications that manage routine and operative issues effectively and efficiently are preferred, not considering employees' thoughts on this matter and their compatibility with AI leads to various concerns.

In summary, while the study investigates whether using AI-based tools in an employee's professional life enhances task performance, it aims to reveal the effects of the level of knowledge about AI, attitude towards AI, trust in AI, and level of discomfort with AI on performance.

The data was collected through the participation of employees working on-site, remotely, and in a hybrid model in one of Turkey's leading corporate firms headquartered in Istanbul. 106 employees participated in the study, but due to inconsistencies in the responses, only 96 participants were included in the analysis. The participants consisted of individuals actively using the Artificial Intelligence-based application of the relevant company and able to provide feedback on it. Reliability analysis, Exploratory Factor Analysis, and Model Test were used to analyze the data in investigating the hypotheses created, and the outputs were examined separately. The findings were elaborated and presented to the reader in detail to prepare a groundwork for future studies.