Carolina A. de Lima Salge bio photo

Carolina A. de Lima Salge

Algorithms.
Networks.
Ethics.

Email Twitter LinkedIn GitHub Scholar

Major Research

I have three major areas of research (each discussed below) and I am adept at a variety of methods, including the case study approach, meta-analysis (quantitative & qualitative), structural equation modeling, and applied econometrics. In the future, I expect to utilize more experiments (both in the lab and in the field).

Bots and Information Dissemination in Online Social Networks

Bots are computer algorithms in online social networks (e.g., Facebook and Twitter). Public interest in bots has exploded in the last two years, as evidenced by the increased attention bots have received in academia and the popular press. One likely reason for such increased attention is that bots are being used to spread low credibility information on social media, and manipulate political processes around the world. In response, many bot detection algorithms have developed (e.g., Botometer) in parallel with some empirical studies on information diffusion by bots. The goal of this stream of research is two-fold. First, to develop "bot theory" by explaining the processes by which bots disseminate information in online social networks. Second, to provide holistic and theory-based procedures researchers can use to better detect bots and understand their actions.

  • Salge, C. & Karahanna, E. 2018. Protesting Corruption on Twitter: Is It a Bot or Is It a Person? Academy of Management Discoveries, 4(1), 32-49. [Journal Link]
  • Salge, C., Karahanna, E., & Thatcher, J. [Title Redacted]. Second Round Revision at MIS Quarterly.

Conversational Agents in an Increasingly Hyper-Private Web Browsing Environment

Conversational agents (also known as chatbots) are natural language computer programs designed to approximate human speech (written or oral) and interact with people via a digital interface. Although they have existed since the sixties (e.g., ELIZA), conversational agents have recently garnered industry attention. They are becoming the new front-office face of many companies, representing a shift from clicks to conversations and from e-commerce to conversational-commerce. Gartner (Juniper Research) predicts that by 2020, 85% of all customer interactions will be handled by conversational agents resulting in cost savings of $8B by 2022.

At the same time that conversational agents are becoming more popular, especially for data collection purposes, the Surface Web is becoming darker and more private. States are passing privacy acts that protect consumers from entities that collect and misuse their personal information by granting them the right of privacy (e.g., see California’s AB 375 bill). In so doing, they are raising awareness about the costs and risks of freely sharing information with firms (and their conversational agents) and thus exacerbating the concerns consumers already have. The objective of this stream of research is to address these opposing trends and generate theoretical and practical insights on the use of conversational agents in areas such as finance, healthcare, and customer service, among others.

  • Thomaz, F., Salge, C., Karahanna, E., & Hulland, J. [Title Redacted]. First Round Revision at Journal of the Academy of Marketing Science.
  • Salge, C. & Karahanna, E. A Systematic Literature Review on Conversational Agents: What Do We Know, Don’t Know, and How to Move Forward? Target: Journal of the AIS.

Algorithmic Transparency and Ethics

People are being constantly observed by computer algorithms. Today, smart cameras and wearable tracking devices make possible a kind of real-time “Super-Vision” far beyond any level of human observability envisioned by Taylor in Scientific Management and passionately aspired by Sting and The Police in Every Breath You Take. These computer algorithms, viewed as recent AI developments, represent an evolution of observability in management. Observability of such computer algorithms, however, is rare. This opacity and lack of visibility has fueled a call for more transparency, and especially more algorithmic transparency. While the ability to observe computer algorithms can enable the judgment of their ethics, the ethics of computer algorithms is in itself very complex partly due to the random or unknown nature of the logic underlying many computer algorithms. The goal of this stream of research is to disentangle some of this complexity by clearly describing what is unique about the "transparency" and "ethics" of the "computer algorithm" and by theorizing on such uniqueness(es). Potential practical outcomes of such theorization include readily accessible guidance, in the form of evaluation procedures, conceptual models, or policy recommendations, rooted in sound ethical thinking.

  • Salge, C. & Berente, N. 2017. Is That Social Bot Behaving Unethically? Communications of the ACM: Vol. 60, Issue 9. [Journal Link]
  • Salge, C., Tumbas, S., Berente, N. & Jetha, K. Managing AI: How "The Right Level" of Algorithmic Transparency Enhances Organizational Value and Mitigates Privacy Concerns. Target: MIS Quarterly.

Other Research

In addition to the above research streams, I also have projects in areas such as sports (mostly tennis and football), escalation of commitment, online communities, and virology (avian influenza vaccines).

  • Berente, N., Salge, C., Mallampalli, V. K., & Park, K. Escalation and Rationality: Logics of Legitimacy and Project Persistence. Target: Organization Science.
  • Salge, C., Goodhue, D. L., & Thompson, R. Exploring the “Evolution-in-Use” of Technology: A Sensemaking Perspective. Target: MIS Quarterly.
  • Salge, C., Berente, N., Howison, J. Participation and Knowledge Sharing in Virtual Communities: A Meta-Analysis. Target: Organization Science.
  • Lisa, G., Montague, N., Cohen, J., Salge, C., & Wayne, J. [Title Redacted]. First Round Revision at AUDITING: A Journal of Practice & Theory.
  • Silva, M. S., Criado, M. F., Lee, D-H, Salge, C., Spackman, E., Donis, R., Wan, H., & Swayne, D. E. Protection by Inactivated H5 Vaccine Strains Against Goose/Guangdong Lineage H5N1 Highly Pathogenic Avian Influenza Viruses in Chicken Model. Target: Journal of Virology.
  • Byon, K. K., Salge, C. & Baker, T. A. [Title Redacted] Conditional Acceptance at Sport Marketing Quarterly.
  • Salge, C., Byon, K. K., & Baker, T. A. 2015. The Limiting Use of Meta-Analysis in Sport Management: A Case of Constraints and Sport Leisure Consumption. Journal of Contemporary Athletics: Vol. 9, Issue 2. [Journal Link]
  • Salge, C. 2010. The ATP World Tour: How Do Prize Structure and Game Format Affect the Outcome of a Match? [Master Thesis]