Comprehensive examination of the bright and dark sides of generative AI services: A mixed-methods approach☆

被引:1
作者
Yoon, Sang-Hyeak [1 ]
Yang, Sung-Byung [2 ]
Lee, So-Hyun [3 ]
机构
[1] Dongguk Univ, Coll Business, Dept Management Informat Syst, 30 Pildong Ro,1 Gil, Seoul, South Korea
[2] Kyung Hee Univ, Sch Management, 26 Kyungheedae Ro, Seoul, South Korea
[3] Kyonggi Univ, Dept Ind & Management Engn, 154-42 Gwanggyosan Ro, Suwon, Gyeonggi Do, South Korea
关键词
Generative AI; Valence framework; Mixed-methods approach; Joint sentiment topic modeling; Expert interview; ChatGPT; GUIDELINES; TRUST;
D O I
10.1016/j.elerap.2025.101491
中图分类号
F [经济];
学科分类号
02 ;
摘要
Recent advancements in artificial intelligence (AI), particularly in generative AI (GAI), have significantly influenced society, prompting extensive discussions about their societal impact. While previous research has acknowledged both the benefits and challenges of AI, the rapid development of GAI has often proceeded without sufficient focus on actionable strategies to address potential risks and unintended consequences. Understanding both the positive and negative aspects of GAI is essential to ensure that technological progress is balanced and responsibly managed to mitigate potential risks and societal harm. This study identifies the positive and negative aspects of GAI from both public and expert viewpoints by applying a valence framework. Using a mixed-methods approach that integrates joint sentiment topic (JST) modeling with the combined use of ChatGPT and expert interviews, we investigated the key positive and negative factors associated with GAI. By integrating the insights gained from these different perspectives, the study proposes strategies for the effective and responsible use of GAI. The study contributes to the existing body of knowledge on GAI by offering a comprehensive understanding of its implications and providing guidance for its ethical and appropriate applications.
引用
收藏
页数:13
相关论文
共 63 条
[1]   Individualism and internet addiction: the mediating role of psychological needs [J].
Arpaci, Ibrahim ;
Kesici, Sahin ;
Baloglu, Mustafa .
INTERNET RESEARCH, 2018, 28 (02) :293-314
[2]  
Axelrod Bradley N, 1992, Clin Neuropsychol, V6, P143, DOI 10.1080/13854049208401851
[3]  
Aydin O, 2023, Acad. Platf. J. Eng. Smart Syst, V11, P118, DOI [10.21541/apjess.1293702, DOI 10.21541/APJESS.1293702]
[4]  
Baidoo-Anu D., 2023, Journal of AI, V7, P52, DOI [DOI 10.2139/SSRN.4337484, 10.61969/jai.1337500, DOI 10.61969/JAI.1337500]
[5]  
Baldassarre Maria Teresa, 2023, P 2023 ACM C INFORM, P363, DOI [DOI 10.1145/3582515.3609555, 10.1145/3582515.3609555]
[6]  
Belisle C., 2006, DIGITAL LITERACIES L, P51
[7]   SNOWBALL SAMPLING - PROBLEMS AND TECHNIQUES OF CHAIN REFERRAL SAMPLING [J].
BIERNACKI, P ;
WALDORF, D .
SOCIOLOGICAL METHODS & RESEARCH, 1981, 10 (02) :141-163
[8]   Probabilistic Topic Models [J].
Blei, David M. .
COMMUNICATIONS OF THE ACM, 2012, 55 (04) :77-84
[9]   Using VADER sentiment and SVM for predicting customer response sentiment [J].
Borg, Anton ;
Boldt, Martin .
EXPERT SYSTEMS WITH APPLICATIONS, 2020, 162
[10]  
Boyce C., 2006, Pathfind. Int.