Journal of Management Information Systems

Volume 38 Number 4 2021 pp. 889-892

Editorial Introduction

Zwass, Vladimir

ABSTRACT:

Approaching the fifth decade of its publication, JMIS has always stood for a broad understanding of the remit of our discipline and for the multiplicity of research methodologies that undergird its sociotechnical approach to the study of information systems (IS). Indeed, we have been privileged to contribute significantly to the definition of our boundaries. The recent years, combining a broad societal change with the vastly enhanced role of IS in society whose transformations it is driving in many ways, validate this approach to our field.

The present issue of the Journal contributes to our understanding and, one hopes, action in two crucial realms. The Special Issue focusing on fake news (FN) we present to you offers a comprehensive (insofar as special issues can do that) analysis of a global societal ill enabled by the Internet. To follow it, an emergent section of three papers presents the societal good enabled and to a degree driven by IS: the approaches and the boons of health informatics. As we study the deleterious effects of FN, we should always keep in mind the good that the Internet-Web compound has brought the world in the societal growth and in the economic development.

The Special Issue on Fake News on the Internet, guest edited by Alan R. Dennis, Dennis F. Galletta, and Jane Webster, offers a multidisciplinary analysis of the destructive global phenomenon. The Guest Editors provide a valuable introduction to the issue they have edited to which 80 submissions were received. It is my place here to add some context and perspective.

“Fake News” is a shorthand for the production, dissemination, and amplification of misinformation originated with an intent to deceive. On the Internet, it may be a fake product review, placed by a competitor, and conceivably integrated into a hostile campaign by a firm specializing in such services. It may also be fake information about a political candidate deployed as a component in a political campaign devised to pervert the democratic process of an election. The first FN may damage a brand, the second may distort the outcome of the election and, beyond that, shake people’s faith in democracy. The actors may be individuals or it may be the Russian Internet Research Agency specializing in disinformation as a state organization formed for influence and propaganda purposes. In fact, state actors may be joined by unaffiliated individuals to diffuse and obscure responsibility, and to provide deniability.

FN is not a novel phenomenon. The arrival of new media technologies always has dual, Janus-like, effects, with the dark side being the reverse of the bright coin. The invention of the printed press in the 15th century did not only bring access to the Bible and literacy to the now more enlightened populace. For centuries, it has also brought threats and transformation to the established orders. Jean-Jacques Rousseau, a philosopher who was no particular friend of his contemporary order, had this to say in 1750 for the ages: “If we consider the horrible disorders that printing press has already produced in Europe … ” (cited in Rauch [6], p.120). Scurrilous printed pamphlets bearing salacious FN about the respective royal consorts kindled the flames of two revolutions, French and Russian, by delegitimizing and “de-majestifying” the reigning houses. Some FN books have lasted centuries, in the ever new editions.

However, there is a huge difference. Powerful technologies have powerful consequences. The power of the Internet and associated information technologies lends the online FN a particular weight. Social media are the principal means of the unholy spiral in spreading FN. Yet the phenomenon of rapid social amplification on the Internet was identified well before the arrival of social media [10]. Here are the principal reasons the FN on today’s Internet present a threat to social order and its democratic transformation. The public sphere created by the Internet-Web is massive and global. The global and dense network of networks and software-stack connectivity can import impacts across distant borders. The smartphone as an edge device is accessible to billions of people around the globe. Social media, with the prevailing anonymity of the actors, are based on platforms owned by colossal tech companies that absorb huge profits and massive data. Big data aggregates allow microtargeting of algorithmically personalized real or fake news at the algorithmically defined best responders. Thus, frictionless access to powerful social media platforms facilitates the origination and spreading of FN. Monetization of FN through advertising is a potent driver of their ubiquity. Softbots delivering news to the users according to their profile, magnify the phenomenon. Business models of smaller rogue firms, actually based on exploiting FN, are particularly effective in disseminating online falsehoods at a profit when the costs of communications approach zero.

The onymous author of an article in established print media bears the mantle of expertise, experience, and authority, and her reputation and career prospects are at stake when she publishes. The bearer of FN on the Internet can be just about anyone. And many in the reading public are not equipped to tell the difference and are prone to share and endorse. With that, we have what has been called “a hype machine” [1]. The problem we are dealing with here is that much of what is being hyped is false and there are many falsities that are being hyped.

It is important to recognize and study all these factors in the blooming of FN in order to seek the points and means of intervention and countermeasures. Like the interdisciplinary study that is required, the measures will entail laws and regulations, a variety of technological means, international cooperation, and education. A fundamental issue is that the spread of FN is based on IT and driven algorithmically. To match this, Facebook employs 15,000 moderators to battle misinformation [3] – and that is woefully insufficient if only in matching the speed of the spread (also, who said that the online big tech does not generate jobs?). Any movement towards solutions has to encompass a variety of means, certainly technological and algorithmic, but going well beyond that.

On the demand side, we should recognize the susceptibility to the false belief lent to FN. This is particularly so during the years of the pandemic that bear social isolation, cliquishness, and sense of vulnerability. Of a lie there is an infinite variety; there is only one truth. A lie can be made infinitely colorful; truth attracts the adjective “stark.” Lie then entices attention; attention can be sold. The very nature and the often skillful design of FN bear more novelty or sensationalism than the real information, which makes them attractive algorithmically and otherwise to the platforms supported by advertisement. Conspiracy theories are more exciting to many than straight news. Confirmation bias makes people susceptible to the FN reaffirming their prior beliefs. To reproduce the apt title of one of the empirically grounded publications of our field: “People believe what they want to believe when it makes no sense at all” [4]. The empirics confirm that FN spread faster and more broadly than real news [8]. The virality of lies has been well expressed by Jonathan Swift: “Falsehood flies, and truth comes limping after it” [7].

The FN phenomenon has marked societal effects. The post-factuality we experience damages the epistemic order and open societies. The erosion of trust in a society has been linked by extensive research to the damages to social order and to economic prosperity. Owing to its sociotechnical nature, our discipline has much to offer in seeking out the roots and solutions to the scourge of FN. For example, social norms have been found effective in combating FN on social media [2]. Formal modeling has been deployed to recommend platform policies to counteract FN [5]. Rating the reliability of news sources, labeling individual messages, blocking messages deemed unreliable by various criteria, limiting the number of repostings or “likes” certainly address the issue, but certainly do not scale up to the multifaceted problem. Our research must inform the promulgation of laws and regulations that will necessarily accompany the technological fixes. Over the recent years, JMIS has published a series of papers proposing and testing solutions to some of the deception problems in various contexts, some as early as 2004 [9]. We shall continue to do so.

The papers in the present Special Issue analyze numerous problems associated with the supply of FN and the demand for them, mutually reinforcing actions. One of the papers offers a valuable systematization of the multidisciplinary (as necessary) research on FN and induces an integrated research model. Some of the papers also propose solutions and study empirically their effectiveness. The efforts are valiant and present valuable remedies. We need to stress the obvious: at present there is no comprehensive solution. This is another of the massive problems we are faced with since the Internet has been overlaid by the Web, such as security and privacy. The combination of multiple methodologies, positivist, interpretive, and design-scientific included, is at our disposal. We must persevere, since our discipline has much light to shed in all of these dark areas, and so we hope.

Three papers that follow the Special Issue form an important section of its own, as they bring the IS research to bear on health and wellbeing. The first of these works, authored by Shuo Yu, Yidong Chai, Hsinchun Chen, Randall A. Brown, Scott J. Sherman, and Jay F. Nunamaker, Jr., addresses the use of IT in the remote detection of human falls, potentially life-threatening events, particularly in the independent living of seniors. The paper advances our ability to deploy wearable fall sensors to a greater advantage along several avenues. It avoids manual feature extraction of the sensor-provided data, which is labor-intensive, relatively subjective, and not scalable. The approach taken here combines deep learning with the explainability of the results. The authors propose and test a Hierarchical Attention-based Convolutional Neural Network as the implementation of the approach they take. The work uses design-science methodology to demonstrate the solution’s worth and superiority with benchmarking against large fall-detection data sets and with a case study.

Two subsequent papers apply IT to two crucial problems in a hospital setting. The first of them, by Yu-Kai Lin and Xiao Fang, targets predictive analytics at the adverse events that befall patients in hospitals. Known as iatrogenic events, these occurrences are the consequence of the medical management of a patient rather than of the disease itself, and they are not uncommon. The costs to the patients’ health and wellbeing are high, and so are the financial costs. Within the design-science paradigm, the authors present a novel model for the prediction of in-hospital adverse events which integrates several sophisticated techniques. The empirical evaluation demonstrates that the authors’ model outperforms other known methods. Beyond that, the authors use simulation to show the cost saving that could accrue to the implementation of their approach in the inpatient care. The reduction of adverse events with the use of the proposed model is of an obvious practical value.

The work of Pankush Kalgotra and Ramesh Sharda uses modeling to predict the patient’s length of stay in the hospital at the time of admission. Owing to the many factors involved, this is a difficult problem, complicated by the potential comorbidities that can emerge during the stay, beyond the primary complaint and the comorbidities identified at the time of admission. The need to readmit a patient owing to a curtailed hospital stay obviously does not serve well any of the stakeholders, the patient in particular. To address the problem, the authors define the construct of latent comorbidities, comprising the network of the historical and probable comorbidities and their interdependencies. The researchers use deep learning fed by a very large database of patient records to model the determinants of the length of stay and to predict its duration.

The three health-informatics papers taken together offer a clear and significant benefit to medical practice – and demonstrate the capabilities of our discipline to deal with societal problems.