ABSTRACT:
We shall live with artificial intelligence (AI). Of course, we have used AI in many of our pursuits for decades now—and it has been a significant and effective tool, a moving frontier of the computerized capabilities. Things are changing in our relationship with AI, arguably at an accelerating speed. The mutually reinforcing information technologies (IT) are cumulatively altering the way we live at work, at leisure, and in the activities combining both. Machine learning based on the deep neural networks, supported by the exponentially growing data aggregates stored and processed in the ever expanding cloud, that is being built the ever more powerful general-purpose and specialized AI chips, places AI ever more closely alongside us humans. The views on these AI developments range from catastrophist to Pollyannaish, both voiced by multiple authorities engaging in responsible futurism. As ever with technologies, a Janusian view—the good to be brought out, with the bad to be identified and neutralized—is justified. The risks, perhaps even existential ones, have to be recognized and managed as early as we can (now) and as well as we can. At the same time, it is our responsibility as humans and specialized scientists, to advance the knowledge of what living with AI means now and will mean in the future. Beyond that, it is our responsibility to use that knowledge to affect the trajectory in the positive direction.
With generative AI (such as GPT-4, as one example), we are moving in the direction of the artificial general intelligence that can challenge human capabilities on a broad front. Moreover, this technology is fungible with other information systems (IS) to make them far more powerful. The distance between the humans and the AI artifacts is still immense, however. And we are the creators of the systems exhibiting these capabilities. This is the time for our discipline to focus our scholarly attention on the AI that is becoming our partner and not only our tool. Very recently, the Journal of Management Information Systems (JMIS) has announced a call for papers to be included in a special issue devoted to this study (https://www.jmis-web.org/cfps/JMIS_SI_CfP_Generative_AI.pdf).
Also recently, JMIS has published a number of papers that apprehend the core issues of our living with the algorithmically-guided systems. Here are the works published just last year, in the 39th volume of JMIS. Chandra et al. [2] study the business impact of the human-like conversational AI agents, to arrive at the desirable competences that would make them attractive to human engagement. Brynjolfsson et al. [1] find the devaluation of human persistence in the collaboration of computerized systems with humans in routine jobs. You et al. [6] provide the theorization and empirics to contrast the human and algorithmic advice in influencing human judgment. Park et al. [4] empirically analyze why the family members of patients reject AI in healthcare. Cram et al. [3] furnish nuanced understanding of the role of algorithmic control in inducing technostress in the setting of the gig economy, where the platform companies are eager to resort to this form of control.
The partnership view of AI is in the lens adopted by the opening paper. Alan R. Dennis, Akshat Lakhival, and Agrim Sachdeva investigate the ramifications of adopting AI agents as members of virtual teams in organizations. How will the other team participants perceive the work-oriented attributes of the AI agents? Will people wish to work with them as team partners? The findings are nuanced, yet in general the answer is to the second question is positive. These findings signify that major changes are upon us that are beyond the major changes we are already seeing in the virtualization of work and of the workplace. We will need to further investigate redesigning our individual and collaborative work in seeking increased productivity from the human-AI collaboration. We will also need to see to it that this growth benefits all members of our organizations—and of our society. This work is generative, as we seek to understand more broadly the roles “digital employees” can play in organizations and beyond. Indeed, the first author has recently collaborated in categorizing these roles and outlining their functionalities [5].
Organizations seek adaptability and flexibility, particularly in their readiness to rapidly respond to the challenging events in their environment. Clearly, those are aims rather than accomplished facts. However, counteraction to organizational rigidities should be a part of this continuous struggle. The authors of the next paper, Jing Gong, Yi Liang, and Narayan Ramasubbu, find empirically and with a theoretical support, that the diversification of the vendors supplying the infrastructure software to the firm is a source of organizational rigidity. The authors find the grounding for their thesis in their study of the performance of close to a thousand large firms during two environmental shocks. These findings are weighty in identifying a general threat factor that may go unnoticed as our software bases grow.
The ownership of consumers’ (read everyone’s) personal data is a hotly contested field. The business models of major technology companies rely on this ownership while the privacy of individuals as members of a well-ordered society hinges on a level of attenuation of the present data-ownership regime in favor of the individual’s control of the coded facts of their life. In the exchange economy, where data can generate revenue, the data becomes property and the allocation of rights to this property calls forth an extensive investigation in the economic and legal terms. Here, Shilei Li, Yang Liu, and Juan Feng present a formal model of the effects of various degrees of mixed data ownership on the mix participants: service provider collecting the data and consumer acting, in effect, as data supplier. With variables such as consumer’s sensitivity to data ownership and the suppliers’ ability to provide differentiated quality of service to compensate for the cost of the data, the authors surface several conclusions that make an important contribution to the data-ownership debates and the role that needs to be assigned to the regulators.
Online games are becoming the ever more important component of the Internet economy and the economy at large, far beyond their entertainment factor. On the one hand, such games are one of the foundational technologies of the emerging metaverse; on the other hand, they enhance the engagement of the participants in various organizational and learning environments. The outcomes achieved with social gamification are the object of an extensive study by Jun Zhang, Qiqi Jiang, Wenping Zhang, Lele Kang, Paul Benjamin Lowry, and Xiong Zhang, presented in the next paper. The findings come from different settings and include a variety of cooperation and competition effects that are valuable to our appreciation of the role that social gamification can play. We need to look further into the effects such games can play in societal bonding and conflict resolution, which are appreciable and worth pursuing.
Several following papers investigate the online behaviors of individuals and collectives in various contexts. Yang Pan, Sunil Mithas, J.J. Hsieh, and Chewei Liu address online trading. The authors investigate empirically the relationship between the intensity of the online channel use by the investors and their trading behavior, with the consideration of their risk acceptance. Grounding themselves on a large data set, the authors produce differential findings on the performance of the risk-accepting and risk-avoiding investors. Notably, these findings help in predicting the investors’ performance based on their use of online trading channels.
Related to the foregoing paper are the findings presented in the next work, by Wael Jabr, Abhijeet Ghoshal, Yichen Cheng, and Paul Pavlou. The authors are seeking to understand and predict the customers’ revisiting of online retailers and purchasing from them. This approach goes well beyond modeling a traditionally explored visit-purchase scenario. A more realistic and more often enacted sequence is of a potential customer visiting the online site once or more and then—perhaps—making a purchase. Subsequently, a relationship may be established between the customer and the vendor, based on a positive experience and trust. The authors develop a two-stage model of revisiting and purchasing, with a consideration of the customers’ heterogeneity. They proceed to validate their model with clickstream data and to establish its superiority with respect to the existing purchase-prediction methods. The work contributes significantly to our understanding of the consumer behavior in e-commerce and to its revenue-seeking practice.
Social media are promising fishing grounds. Fishing has been growing apace and this type of cyberattack leads in the aggregate to massive losses, often affecting those who can least afford it. The individual susceptibility to phishing on social media is the subject of the work by Hamed Qahri-Saremi and Ofir Turel. The authors present their validated temptation-restraint model of individual susceptibility to a fishing attack. The model is couched in the situational contingencies that combine the individuals’ vulnerability in the moment with their perception of the attack’s source and the effectiveness of the fear appeals. The model is comprehensive and opens new vistas in fending off fishing attacks.
Video-sharing sites are a growing online consumption locus, with YouTube a leading brand with very high revenues, based on massive viewership. Obviously, the viewership is not to be taken for granted, as its volume is being challenged by the rising competitors and buffeted by the economic environment. A model offering accurate prediction of viewership is therefore of high value. This is what Jiaheng Xie, Yidong Chai, and Xiao Liu offer in the next paper of the issue. Their model relies on deep learning to both raise the prediction accuracy above the level currently available and to offer the interpretability of the predictors. This a significant contribution to the design research in our discipline, as it is portable to other settings.
The role of the access devices in our cognitive processing is explored by Kathrin Figl and Ulrich Remus. More specifically, they answer the question whether the use of a smartphone makes us think “faster,” in the sense of more intuitively and less reflectively, than the use of a PC or a similar large-screen device. The work is carried out empirically while the authors ground themselves in our knowledge of the functioning and capacity of working memory. The answer is: “It’s not in the device you use, it’s in you choosing the device.” This answer belies the results reported in several previous studies and has general weight.
To guide them in their selection of products and services, consumers use the platforms that derive their aggregated information from the crowdsourced data, in other words, from other consumers. Often called online business directories, such platforms (say, Yelp or Tripadvisor) facilitate decision-making and are commonly used before a transaction is made. By their very nature, these data aggregates are often incomplete, depending on the volunteer behavior of the individual explicit or implicit data suppliers. Da Xu, Paul Jen-Hwa Hu, and Xiao Fang present an imputation method to derive the missing attribute values from the data available in the directories. As a preceding paper in the issue, the work relies on deep learning and contributes to our design-research body of knowledge.
The following paper also relies on machine learning and offers a design-research contribution, albeit in a different sphere of application. Ruiyun Xu, Hailiang Chen, and J. Leon Zhao present a novel method and a system that enhances our ability to offer disciplined recommendations for venture-capital funding. The theory-based SocioLink framework the authors develop and prototype matches the investor and the—potential or existing—startup via a knowledge graph reflecting different connections among the players. The model offers the always desired explainability of its recommendations, and the authors show that their approach outperforms the existing methods of venture-capital allocation. The work will undoubtedly be continued to avoid unduly privileging the existing relationships and to enrich the knowledge graph with new potentialities.
The concluding paper of the issue investigates the behavior of crypto tokens on blockchain platforms. Here, crypto tokens are used as rewards to encourage user participation. They are not money, but they have a price. That price usually swings widely. The researchers, Kun Chen, Yifan Fan, and Shaoyi Stephen Liao investigate how the reward uncertainty owing to this volatility affects users’ contribution in a tokenized digital platform. The setting of their empirics is a platform where bloggers write posts and obtain token rewards based on their posts’ popularity. The authors find that high token-price volatility induces a large volume of posts but diminishes their quality. The authors’ analysis further shows that token price volatility has a positive effect in the short-term yet impairs long-term creativity of bloggers. The work contributes significantly to our understating of motivation and incentives in the expanding blockchain environment, and beyond.