Professor Sangseok You interviews on the publication of his research paper Journal of Management Information Systems
- bizskk
- Hit563
- 2022-09-07
Q1. Hello, Professor Sangseok You, please introduce yourself and the research field.
Hello, I'm Sangseok You, majoring in business information. In a business environment, I am thinking and researching on the use of intelligent agents, including artificial intelligence and robots, to increase work performance and effective collaboration among employees through such technologies. Recently, I am studying ways to improve customer service using intelligent robots and to increase transparency and reliability in the decision-making process using algorithms.
Q2. Is there any reason why you became interested in the paper you researched this time?
This paper was published in the Journal of Management Information Systems as “Algorithmic versus Human Advice: Does Presenting Prediction Performance Matter for Algorithm Appreciation?”. In this paper, I investigated the mechanism in which the results recommended by the algorithm are understood and trusted by the user.
I was very interested in the fact that the recent rapidly emerging algorithm and recommendation system using artificial intelligence have a growing impact on users' decision-making. Also, I became interested in this field because my other research field, collaboration with robots, is mainly related to physical labor, whereas collaboration with algorithms and artificial intelligence is deeply related to knowledge labor.
Q3. Please introduce the paper in detail.
In this paper, we focus on a phenomenon called algorithm application, in which users comprehensively prefer algorithm recommendations to themselves or others. To get a little closer to this comprehensive algorithmic preference, I and my co-authors envisioned a theoretical framework and demonstrated it empirically in five experiments. It is a theoretical framework that results in the final judgement considering the recommender, the user, and the user's algorithm recommendations, and I hope it will help in future decision-making research using algorithms.
<A Theoretical Framework for Algorithm-Based Decision Making presented in Figure 1>
As mentioned earlier, in five experiments, various variables such as the way the algorithm visually presents recommendations, reliability of the algorithm, and user's cognitive load and temperament cognitive needs were studied. For example, if an algorithm reveals that there is an error in the recommendation process, it is found that user's dependence on algorithm recommendation decreases. In particular, disclosure of summary information without disclosing details of recommendation errors did not significantly affect the decrease in reliability, but disclosure of details of the error resulted in a greater decrease in reliability. Interestingly, users were less affected by human recommendation error information, while more sensitive to algorithmic errors.
<Figure 2 Changes in user's dependence on recommendation according to algorithm recommendation error information disclosure method>
(No Err: Nondisclosure, AAE-AHR: Summary disclosure, Table-Diff: Detailed error content disclosure)
Q4. What was the difficult part of your research?
It may be difficult for all researchers, but it took a lot of time and effort to design and carry out five experiments. In particular, there have been many trials and errors and corrections in technical processes, such as selecting and training data and algorithms for experiments, and in the logical processes of creating a valid experimental environment to effectively prove research hypotheses.
Also, the process of writing a paper was never easy. It was also difficult to easily explain the complex experimental design, but the process of unraveling the importance and theoretical implications of the research problem in clear language was also a part of the researchers' efforts.
Q5. Please tell us about your future plans related to research activities.
In the future, I plan to continue research related to decision making using algorithms and artificial intelligence. In addition, I will continue to develop research on cooperation using robots equipped with artificial intelligence. It is one of my wishes to study with students who are interested in this field with passion and ability.