Research Summary We investigate how firm performance feedback and economic conditions (upturns and downturns) impact changes in the breadth of partner types within a firm's R&D alliance portfolio (R&D AP breadth). We theorize that economic conditions moderate firms' motivation to change R&D AP breadth in response to social and historical performance-aspiration discrepancies, and shape whether they respond more strongly to social or historical aspirations. Using cross-industry data on Swiss firms (1999-2020), we find that firms performing below social aspiration are more likely to expand R&D AP breadth during downturns but less likely to do so during upturns. Performance below historical aspiration shows no significant effect once economic conditions are considered.Managerial Summary We examine how firms adjust their range of business partners in response to past performance and economic conditions. Analyzing Swiss firms from 1999 to 2020, we find that firms strategically adjust the number of partnership types in response to economic climates. During tough economic times, firms tend to broaden their R&D partnership types when they fall short of peer performance benchmarks. They tend not to do so in prosperous periods under similar performance conditions. We found no consistent effects when performance is benchmarked against the firm's history. These insights help managers strategically navigate alliances under fluctuating economic conditions and underscore the importance of peer-relative performance benchmarks.

This article focuses on the deployment of proprietary artificial intelligence (AI) systems in strategy creation and implementation processes, with a specific focus on their role in enhancing organizational attentional control. By employing the attention-based view (ABV) as an overarching theoretical framework, we examine how company-specific AI systems trained on proprietary data can support strategy processes. We identify three key contributions of AI in strategy creation and implementation processes: (1) broadening organizational attention to external and internal developments, (2) democratizing strategic processes through improved transparency and inclusivity, and (3) accelerating feedback loops with real-time monitoring of strategy implementation progress. Potential challenges associated with the deployment of AI systems for attentional control are also addressed. The paper concludes by putting forward potential directions for future research.

We examine how firms adjusted their open innovation strategy in response to the 2008 global financial crisis. While previous research has analysed the advantages, drawbacks, and methods of open innovation, less is known about how firms adjust their open innovation strategy in response to major economic shocks. Guided by theories of organisational learning and behavioural theory of the firm, we examine the impact of demand shock on firm openness to external knowledge. To test our hypotheses, we analyse a unique dataset on innovation in Swiss firms during the financial crisis. Our findings show that firms persisted with open innovation during the crisis, but the nature of the shock had a differential effect on how firms searched for external knowledge. This research contributes to a better understanding of the role of open innovation in times of crisis and provides insights into how firms adjust their innovation strategies in response to economic shocks.

Artificial intelligence (AI) applications have proliferated, garnering significant interest among information systems (IS) scholars. AI-powered analytics, promising effective and low-cost decision augmentation, has become a ubiquitous aspect of contemporary organisations. Unlike traditional decision support systems (DSS) designed to support decisionmakers with fixed decision rules and models that often generate stable outcomes and rely on human agentic primacy, AI systems learn, adapt, and act autonomously, demanding recognition of IS agency within AI-augmented decision making (AIADM) systems. Given this fundamental shift in DSS; its influence on autonomy, responsibility, and accountability in decision making within organisations; the increasing regulatory and ethical concerns about AI use; and the corresponding risks of stochastic outputs, the extrapolation of prescriptive design knowledge from conventional DSS to AIADM is problematic. Hence, novel design principles incorporating contextual idiosyncrasies and practice-based domain knowledge are needed to overcome unprecedented challenges when adopting AIADM. To this end, we conduct an action design research (ADR) study within an e-commerce company specialising in producing and selling clothing. We develop an AIADM system to support marketing, consumer engagement, and product design decisions. Our work contributes to theory and practice with a set of actionable design principles to guide AIADM system design and deployment.

Generative AI is upon us and is changing organizations and organizing. In this essay, we extend the relational perspective on technology, which argues for moving away from an entity-based view of technology to one that focuses instead on the evolving relations and functions between people, technologies, and organizations. We do so by introducing the concept of “progressive encapsulation” which captures GenAI’s potential ability to increasingly expand the “black box” and reduce visibility into and control over the relations and functions performed. We argue that progressive encapsulation is critical in our theorizing about GenAI. As an illustration and thought experiment we consider how GenAI and progressive encapsulation may necessitate changes in our theorizing about groups and teams in organizations.

The growth of digital platforms has led to the proliferation of Online Communities, providing individuals with opportunities to seek help and share knowledge. A key challenge of help-related platforms that address technical questions (i.e., utilitarian, rather than opinion or supportive) is to ensure the contributions address seekers’ specific information needs. Despite growing academic interest in such platforms, research has mainly focused on factors that influence the quantity of contributions, ignoring whether these contributions effectively helped the seekers. To fill this research gap, this study draws upon theories of self-determination and motivation crowding to examine contributing behaviors that result in successful helping. By analyzing a rich dataset collected from an online Q&A platform, we find that gains in a help provider’s past rewards positively influence the success of contribution. Further, while previous studies suggest that external rewards result in a high quantity of contribution, our findings show that an inflated frequency of contribution leads to a crowding-out effect. Specifically, the contribution frequency has a curvilinear relationship with the success of the contribution. Taken together, these findings demonstrate there is a need to revisit the gamification mechanism on help-related platforms to ensure the success of knowledge contribution. This is crucial for the sustainability of these platforms as low-quality answers can lead users to mistrust and eventually leave the platform.

This study investigates the relationship between virtual organizational communication and professional isolation during the change to hybrid work after enforced remote work in a professional services firm. Based on 10 weekly surveys with 1,178 employees, we find that professional isolation increases during the initial stages of the transition to hybrid work. Furthermore, drawing on the technology affordance view, we hypothesize on the impact of virtual interactions with organizational members on professional isolation. We find that employees exposed to virtual senior leadership communication experience a decrease in professional isolation. However, this effect does not hold for virtual team meetings and virtual informal interactions, potentially because these interactions require higher media richness, such as face-to-face contact, to be effective in providing connectedness that reduces professional isolation. Last, we demonstrate that professional isolation impacts hybrid workers’ job performance and life satisfaction. Overall, this study contributes to the isolation literature by revealing the challenge hybrid workers face in the transition to hybrid work; being virtually highly connected yet still experiencing professional isolation. We further discuss practical and theoretical implications.

Artificial intelligence (AI) drives innovation across society, economies and science. We argue for the importance of building AI technology according to open-source principles to foster accessibility, collaboration, responsibility and interoperability.

Academy of Management JournalVol. 66, No. 2 From the EditorsFree AccessRecognizing and Utilizing Novel Research Opportunities with Artificial IntelligenceGeorg von Krogh, Quinetta Roberson and Marc GruberGeorg von KroghEidgenössische Technische Hochschule Zürich (ETH Zurich), Quinetta RobersonMichigan State University and Marc GruberEcole Polytechnique Fédérale de Lausanne (EPFL)Published Online:19 Apr 2023https://doi.org/10.5465/amj.2023.4002AboutSectionsPDF/EPUB ToolsDownload CitationsAdd to favoritesTrack Citations ShareShare onFacebookTwitterLinkedInRedditEmail As we are witnessing a fundamental transformation of organizations, societies, and economies through the rapid growth of data and development of digital technology (George, Osinga, Lavie, & Scott, 2016), artificial intelligence (AI) has the potential to transform the management field. With the power to automatize, provide predictions of outcomes, and discover patterns in massive amounts of data (Iansiti & Lakhani, 2020), AI changes many aspects of contemporary organizing, including decision-making, problem-solving, and other processes (Bailey, Faraj, Hinds, Leonardi, & von Krogh, 2022). AI also enables firms with capabilities for offering new products and services, developing new business models, and connecting stakeholders. In line with these developments, AI is not only an interesting phenomenon to study in and around organizations (e.g., Krakowski, Luger, & Raisch, 2022; Tang et al., 2022; Tong, Jia, Luo, & Fang, 2021), but also offers management scholars a wealth of research opportunities in enlarging their methodological toolbox to leverage vast amounts and various types of data (e.g., Choudhury, Allen, & Endres, 2020; Vanneste & Gulati, 2022). In our quest to push scientific boundaries, we encourage authors to explore these opportunities within Academy of Management Journal (AMJ).While AI is a trending and “hot” topic discussed by policy-makers, business leaders, and scientists, understanding when and how machines display intelligent behaviors is at the center of modern AI scholarship (Nilsson, 1998). AI represents a broad field of technologies that display intelligent behaviors, including self-awareness, goal formulation, goal-directed action, reasoning, optimization, learning, and autonomous movements. While many readers will immediately recall the Terminator in the movie by the same name, or the rogue artificial psychopath (HAL) in the movie 2001: A Space Odyssey, which epitomizes the horrific efficiency of AI (Clarke, 1968/2000), computer scientists have long abandoned the idea of building machines that “mimic” human intelligence. Today, research focuses on exploring how aspects of artificial intelligence work in practice. As a subset of AI, machine learning (ML) has become a powerful means of discovering patterns in massive amounts of data and making predictions based on such intelligence (George, Haas, & Pentland, 2014; Hannah, Tidhar, & Eisenhardt, 2021). As management scholars enjoy unprecedented access to numerous new data sources offering vast amounts of data, we believe that ML may be of great value to scholars across the different divisions and interest groups of the Academy of Management. We also believe management research can offer important insights to the study of AI by enlarging the types and scope of data to which we expose ML.The primary purpose of this FTE is to support scholars in taking advantage of research opportunities in ML applications and to offer guidelines for work submitted to AMJ. Specifically, as ML may have broad appeal to management scholars, we discuss the associated benefits and challenges to be considered when engaging in this type of research. We also review recent advancements in AI that may help management scholars leverage the benefits of ML applications while dealing with some of the challenges of these methods. Finally, we offer a typology for conducting ML research in ways that build and test theory and advancing an understanding of management phenomena, consistent with papers published in AMJ.BENEFITS OF ML METHODSOne key benefit of ML to management scholars is its statistical flexibility, or the use of algorithms to fit complex functional forms to a data set (without over-fitting) that can also work well out-of-sample.1 Current ML algorithms fall into three broad categories: (1) reinforcement, (2) unsupervised, and (3) supervised. “Reinforcement learning” comprises algorithms that autonomously learn courses of action to reach a goal within an environment. For example, such algorithms form the backbone of self-driving vehicles that adapt to traffic patterns as they seek to reach a destination, or of advanced chess programs that learn and adapt based on opponents’ moves. “Unsupervised learning” algorithms learn to discover patterns in input data (often called “features”) without any specified target for learning. Such algorithms do not necessarily require annotated and curated data, which makes them very useful for clustering or dimensionality reduction. For example, unsupervised learning algorithms can support a researcher in identifying groups of firms adopting related business models, pursuing similar strategies within an industry (an initial step in strategic group analysis), or offering competing digital services on a platform based on a large and complex feature set. In contrast, “supervised learning” algorithms, which are most common in the social sciences, learn how to map associations between inputs (“features” X) and outputs (“targets” Y). For example, ML can be used with categorical variables to arrange data as they relate to the target, such as classifying credit card transactions as fraudulent or non-fraudulent based on the time and frequency of transactions, location, amount of payment, and other data. For continuous variables, supervised learning algorithms learn to perform complex regression tasks, such as predicting firm performance based on extensive firm and industry data.CHALLENGES OF ML METHODSBecause AI and its subset of ML-based methods involve considerable human effort and management, there are also associated ethical and technical challenges when applying them. First, limited access to certain types of data can challenge AI and ML methods. With a few exceptions, most ML algorithms only work effectively with large amounts of data. However, management scholars must grapple with matching the data to which they have access to data models. Since many firms consider data to be a strategic asset that must be protected through substantial technical and social mechanisms, the extent to which critical data are available may become decisive for the type of questions that drive a research agenda. Researchers may further be constrained by their ability to obtain structured or annotated data for training and evaluating models. While an unsupervised learning approach or manual annotation of a data set may be used to train models, such annotation is prone to errors when data grow. Therefore, access issues continue to create barriers to the use of ML methods and, therefore, to the examination of pertinent phenomena or development of relevant theories.Second, ML algorithms are subject to human biases that can become reinforcing, as they are reproduced in future data sets and models. Researchers are often biased toward selecting algorithms they know rather than those needed to solve the task at hand, which can produce faulty models that are fatal to accurate predictions or pattern discovery. For example, research highlights biases in ML-based candidate selection systems stemming from hidden biases embedded in data from previous candidate pools (Feuerriegel, Shrestha, von Krogh, & Zhang, 2022). Specifically, algorithms trained on White male candidate pools were shown to choose similar candidates and to discriminate against female and ethnic minority candidates. There is also a potential for ML to fall prey to biases resulting from data structures, as the selection of “correct” algorithms depends on the available data and how they are handled. For example, choosing a deep learning model over supervised learning when data are limited can produce faulty predictions. Biases may also derive from performance metrics to evaluate ML models that fail to detect class imbalances and instead, show high performance of target prediction—thus, creating an illusion of model accuracy relative to other models.Third, researchers must consider the costs associated with ML. In particular, massive and variegated data sets often make traditional ML approaches less efficient and pose important constraints for researchers in understanding such data. We also often hear scholars complaining that the intricate functioning of ML algorithms remains opaque. Of course, this perceived challenge can be addressed through education and learning on how algorithms work. However, in many cases, the amount of one’s own training, and of training data needed, makes it economically inefficient to study their detailed functioning. Thus, the required effort to engage in ML methods presents a challenge to researchers who do not regularly utilize this approach.Despite this range of challenges, recent developments in computer science can help management scholars deal with issues of access, bias, and cost. For example, to build powerful learning models, researchers may draw on data to which they have limited access. “Federated learning” is a technique to train models across a variety of data stored on separate servers without exchanging that data (Zhang, Xie, Bai, Yu, Li, & Gao, 2021), such as utilizing data from CCTVs to monitor and predict traffic patterns across several local conditions and without revealing sensitive personal information. Similarly, “transfer learning,” an ML method in which pre-trained models are used as starting points to model a new task (Niu, Liu, Wang, & Song, 2020), is also becoming increasingly important in industries where data protection is crucial, such as defense and pharmaceuticals. To address biases in ML, scientists have begun developing “explainable AI” that can insert explanation and reasoning steps to show researchers the evolution of models while maintaining their predictive accuracy (e.g., Senoner, Netland, & Feuerriegel, 2022). Further, while these explanations often remain limited to predictions for a local area of the data set, researchers are also developing robust methods for integrating them, and, thus, making them universally applicable for the full data set (Lundberg et al., 2020). To address learning curve challenges related to the use of ML methods, scientists have developed “AutoML” systems, which support the choice and evaluation of learning models (He, Zhao, & Chu, 2021). Such systems automatically tune hyperparameters, thereby removing many otherwise complex manual tasks that demand significant research experience. While manually working through iterations in building ML models may be the best way to learn about the opportunities and threats in model selection, AutoML is promising for expanding researchers’ methods portfolio.ML STRATEGIES TO BUILD AND TEST THEORY IN AMJConsidering that the mission of AMJ is to publish empirical research that tests, extends, or builds management theory, strategies to apply ML in management research should consider the state of theorizing about an empirical phenomenon. Within the landscape of management and organizational research, there are four primary ML research strategies that can be distinguished based on the continuity of the phenomenon investigated and its theoretical coherence: (1) predictive selection, (2) predictive refinement, (3) formative discovery, and (4) reductive discovery. Such distinctions are important given that ML predictions can only be built on relatively stable data and should be used for the purposes of advancing our scholarly understanding of a phenomenon. We discuss these four research strategies in greater detail below (see also Table 1).TABLE 1 Machine Learning Strategies for Theoretical Contribution in the Landscape of Management and Organizational ResearchPredictive SelectionManagement research thrives on a proliferation of theories, methods, and data, and researchers often formulate competing features and hypotheses to explain a phenomenon. In predictive selection, data about a phenomenon are continuous and therefore offer an opportunity to use ML to predict outcomes. At the same time, there is theoretical fragmentation in terms of the availability of many (and maybe competing) priors, such as candidate theories and variables, for explaining the phenomenon. Thus, ML can be used to examine the predictive strength of these priors (features) relative to a dependent variable (target). For example, He, Puranam, Shrestha, and von Krogh (2020) studied how open-source software projects resolve collective disputes around the choice of intellectual property rights. Framing their study in a “governance of the commons” perspective (Ostrom, 1990), the authors argued there is limited theoretical explanation of how communities resolve disputes over principles to license their open-source software. After manually identifying 11 priors from existing theories and observations on structural and processual features of governance disputes, they exploited these priors to code the entire sample of disputes (183 projects). Then, after building an ensemble of ML algorithms to detect robust association between those independent variables and dispute resolution outcomes (resolved or non-resolved), four algorithms were run and evaluated to achieve the best fit. The model demonstrated that the size of the group involved in a dispute, active efforts to add information in discussion, application of a choice procedure (e.g., voting), and type of issue under dispute (e.g., changing an existing or selecting a new license) predicted the resolution outcome. Based on these insights, the authors created a subsample of 61 cases that guaranteed minimum and maximum variance in observations to “manually” build a theory of governance dispute resolution. Thus, this study demonstrates how ML can be useful for inductively building or elaborating theory.Predictive RefinementML methods can aid researchers further when the stability of the phenomenon enables prediction by ML. Predictive refinement is a viable strategy when alternative theories about mechanisms explaining the phenomenon have become gradually integrated, and the variables and measures are relatively well established. For example, Rathje and Katila (2021) examined firms’ investment decisions and the impact on the enabling nature of technologies using a data set comprising the full set of patents awarded to U.S. firms between 1982 and 2022 (almost two million patents). Predicting antecedents of relationship (e.g., grants) and the effects of partner type (e.g., science agencies), the authors uncovered differences across firms inventing such technologies alone versus with public organization partners, with a subset of 33,130 patents resulting from private–public R&D relationships. Rathje and Katila (2021) conducted a quasi-experimental design drawing upon supervised ML to build and investigate the treatment versus control groups. ML was used to perform propensity score matching, which helped to identify balanced subsamples of patents with respect to observed covariates. This learning model avoided intractability associated with conventional propensity score-matching methods and enabled the authors to consider a much larger set of potentially confounding dimensions. In particular, regularization and cross-validation made it possible to circumvent the issue of overfitting of the model to the data, and the ML training sample, optimization techniques, and holdout sample were used to make the most accurate predictions. Thus, this approach illustrates how ML can be used to fine-tune causal inference, perform sensitivity analyses, and refine measures with an ultimate goal of refining theory.Formative DiscoveryA researcher may be confronted with a discontinuous phenomenon wherein the data set does not allow for accurate predictions through ML. Additionally, theory about the phenomenon may be fragmented. Under these conditions, ML can support formative discovery identifying patterns in the data set useful for gaining insights into the phenomenon. Specifically, formative discovery is a strategy that can help scholars build unprecedented and often pre-theoretical understanding of a phenomenon by revealing patterns that may not be obvious to even the best-trained observer. For example, drawing on a rapidly emerging and foundational literature on the outcomes of diversity in organizations, Wang, Dinh, Jones, Upadhyay, and Yang (2023) studied alignment between firms’ diversity statements and their employees’ online rating of those firms’ diversity, equality, and inclusion (DEI) efforts. As several firms publicly condemned racism and affirmed their stance on DEI following the rise of the Black Lives Matter social movement and the death of many Black Americans during the spring of 2020, the authors drew attention to a limited scholarly understanding of the contents of such statements and their impact on organizational outcomes. To explore the statement–outcome link, the authors first applied structural topic modeling (STM), an unsupervised ML technique, to data from open letters to stakeholders across the Fortune 1000 during late May to early June 2020 to identify themes (patterns) in DEI statements. After training, the learning models identified general DEI themes related to supporting and acknowledging the Black community and committing to diversifying the workforce, which were used in a second study targeting millions of data points of employees’ DEI ratings of firms on Glassdoor.com. Through the performance of topic probability scores (based on STM) for each firm’s statement, which represent the likelihood that a statement falls into a particular theme, the results showed that firms that released public statements and referenced identity-conscious topics received more favorable DEI ratings from their employees. Accordingly, this study is an exemplar of how ML can help to discover patterns in massive data sets, and thereby provide novel conceptual insights into poorly understood phenomena.Reductive DiscoveryResearchers are often challenged by a need to identify limitations to existing theory. Suitably, reductive discovery strategies are useful for identifying patterns in data that show such conceptual limits to generalizability. While limits to generalizability may be revealed through replication studies, there is additional value in using ML to examine emerging patterns in novel data sets on a discontinuous phenomenon. ML can reveal hitherto unknown associations across features, bringing into question the applicability of existing theoretical explanations, and offer up novel learning opportunities. For example, Belikov, Rzhetsky, and Evans (2022) developed a promising approach to mitigating the challenges of generalizability using Bayesian calculus to predict robust scientific claims based on data extracted from prior publications and weighted by institutional, social, and scientific factors. The authors applied ML to the case of gene regulatory interaction studies and identified a set of fundamental characteristics that predict replicability across contexts, which have been used to guide scholars’ choices in research topics and programs. While our field thrives on great diversity in categories, constructs, theories, methods and data compared with the life sciences, this diversity is an opportunity for our field, bearing in mind that ML improves task performance over time as the underlying models become exposed to increasingly diverse data.ML STRATEGIES TO BUILD AND TEST THEORY: AN EXAMPLEEngaging in rigorous ML research requires an understanding of the requisite methodological steps as well as the expertise and efforts needed at each step. Because such requirements are apparent in research that makes predictions with supervised learning, we offer an example that highlights relevant choices and outcomes involved in ML methods (Shrestha, He, Puranam, & von Krogh, 2020; Shrestha, Krishna, & von Krogh, 2021). Specifically, because an ML model (as a specified set of computations built as ML algorithms) works through different values in the data when solving a task, we discuss the three steps in building a learning model: (1) data management, (2) learning, and (3) evaluation.During data management, the researcher specifies the task to be completed and the relevant data sources, and proceeds to collect and analyze the data (identifying distributions, skewness in data, or class imbalance). She then divides the data into two sets—the “training data,” used to iteratively improve the model, and the “holdout data,” used to evaluate the predictive power of the model (80:20 or 70:30)—and applies techniques to deal with outlier treatment and missing values. The researcher also conducts “feature engineering,” which specifies features in the data used to build the model. Feature engineering is constructing, deriving, and transforming data into variables that can be used as input features for the model (Chapman, Gilmore, Chapman, Mehrubeoglu, & Mittelstet, 2020). For numerical data, this may imply normalizing the data by features into a 0:1 interval. Accordingly, a sound ML design relies on finding the right features on which to train the model and reporting on the details of this activity. It is useful to sense-test the features with the existing literature, since the purpose of ML is to help researchers build and test theory.During the learning stage, the researcher chooses a performance metric for the prediction task (e.g., log loss score). Once the correct performance metrics have been established, ML algorithms must be chosen to fit data type and data size, prediction accuracy, interpretability of the model, extent of the feature set, etc. Notably, there are many specialized ML algorithms from which to choose for classification (Y is a categorical variable) or regression (Y is a continuous variable) tasks. Some will be familiar to management scholars, such as linear and logistic regression, although many useful algorithms are less common and demand careful examination and explanation prior to use (e.g., tree models: decision tree, random forest, gradient-boosted trees). Having chosen an algorithm, the researcher next selects and optimizes the model by searching for hyperparameters (e.g., number of trees, tree depth, batch size, learning rate), which are intended to control the learning process. While some simple algorithms require no hyperparameters, many demand researchers to carefully select such parameters (or automatize such selection) to achieve a good learning performance of the model. This “tuning” is commonly done by applying standard search techniques, such as simple or greedy search (i.e., locally optimal choices at each stage of search) to search for the best fit. Thus, the selection of ML algorithms, as well as the inclusion and tuning of hyperparameters, is critically important in applying ML.During evaluation, the researcher examines how the chosen model performs predictions on the holdout data relative to the set performance metrics. This step involves the analysis of model errors and aims to give the researcher a sense of a model’s generalizability beyond the training data. In addition, reporting the outcome of the evaluation step is necessary for future research to advance on the same or similar data sets. An important aspect of evaluation is also to check usability and deficiency of the model in a real-world context.A CALL TO ACTIONWith a mission of publishing empirical research that significantly advances management theory and contributes to management practice, there are several considerations for authors targeting AMJ as an outlet for their significant value-added contributions to the field’s understanding of an issue or topic via AI and ML. While the Journal has augmented its reviewer capacity and methodological expertise to handle manuscripts with ML applications, there are a few deliberations that will improve the likelihood that a submission will be favorably received. Beyond being clear about why they use ML to discover patterns and build and test theory, researchers should focus on data engagement, data treatment, and data management.Submissions to AMJ that draw on ML methods should demonstrate deep understanding of and engagement with the data. Researchers should take steps to comprehensively familiarize themselves with the data at hand and to ask what can or should be learnt from this data. In so doing, they should pay particular attention to feature engineering—specifying what features can be extracted from the raw data relevant for predicting the target.Authors drawing on AI and ML methods should also consider parameter choices and reflect on how bias may influence those choices and interpretation of their findings. For example, when choosing a model, it is imperative that researchers understand the goals they are trying to achieve. Similarly, researchers should make sure their training data sets are the best representative of the whole population. It is also important that we devote space (in text or in online appendices) to explain such choices and other aspects of feature engineering in detail. It is also important for researchers to share their code (at least during the review process), and we generally encourage publishing code on publicly accessible repositories.Finally, authors of ML-based studies must take particular care in the management of their data. In addition to following ethical standards set by the Academy of Management, scholars who apply these methods to private and publicly available individual data must safeguard data security and privacy. For example, authors should anonymize data and develop a comprehensive data management plan to protect data and avoid any breach of privacy.CONCLUSIONAI should not ossify a creative mind but instead embolden and inspire researchers to engage with novel and unconventional data in their area of interest. The mindset driving this engagement should be: “How can these tools help me uncover what I always wanted to know, but never had the power to study?” Once clarity exists on that question, we may approach our research questions and phenomena with a portfolio of research strategies, including formative and reductive discovery, predictive selection, and predictive refinement. In doing so, we have both the opportunity and the ability to transform management research and practice.1 While classical statistics focuses on extracting inferences for a population from a sample, ML aims at finding generalizable predictive patterns (Bzdok, Altman, & Krzywinski, 2018). This essential feature may support researchers in scrutinizing effect sizes in very large samples (see Combs, 2010).AcknowledgmentsWe are very grateful for comments on earlier versions of this editorial from Savindu Herath, Riitta Katila, Phanish Puranam, and Yash Raj Shrestha. Any errors are our own.REFERENCESBailey, D. E., Faraj, S., Hinds, P. J., Leonardi, P. M., & von Krogh, G. 2022. We are all theorists of technology now: A relational perspective on technology and organizing. Organization Science, 33: 1–18. Google ScholarBelikov, A. V., Rzhetsky, A., & Evans, J. 2022. Prediction of robust scientific facts from literature. Nature Machine Intelligence, 4: 445–454. Google ScholarBzdok, D., Altman, N., & Krzywinski, M. 2018. Statistics versus machine learning. Nature Methods, 15: 233–234. Google ScholarChapman, K. W., Gilmore, T. E., Chapman, C. D., Mehrubeoglu, M., & Mittelstet, A. R. 2020. Camera-based water stage and discharge prediction with machine learning. Hydrology and Earth System Sciences Discussion. Forthcoming. Google ScholarChoudhury, P., Allen, R. T., & Endres, M. H. 2020. Machine learning for pattern discovery in management research. Strategic Management Journal, 42: 30–57. Google ScholarClarke, A. C. 2000. 2001: A space odyssey. New York, NY: ROC. (Original work published 1968) Google ScholarCombs, J. G. 2010. Big samples and small effects: Let’s not trade rigor and relevance for power. Academy of Management Journal, 53: 9–18.Link , Google ScholarFeuerriegel, S., Shrestha, Y. R., von Krogh, G., & Zhang, C. 2022. Bringing artificial intelligence to business management. Nature Machine Intelligence, 4: 611–613. Google ScholarGeorge, G., Haas, M., & Pentland, A. 2014. Big data and management. Academy of Management Journal, 57: 321–326.Link , Google ScholarGeorge, G., Osinga, E. C., Lavie, D., & Scott, B. A. 2016. From the editors: Big data and data science methods for management research. Academy of Management Journal, 59: 1493–1507.Link , Google ScholarHannah, D. P., Tidhar, R., & Eisenhardt, K. M. 2021. Analytic models in strategy, organizations, and management research: A guide for consumers. Strategic Management Journal, 42: 329–360. Google ScholarHe, V. F., Puranam, P., Shrestha, Y. R., & von Krogh, G. 2020. Resolving governance disputes in communities: A study of software license decisions. Strategic Management Journal, 41: 1837–1868. Google ScholarHe, X., Zhao, K., & Chu, X. 2021. AutoML: A survey of the state-of-the art. Knowledge-Based Systems, 212: 106622. Google ScholarIansiti, M., & Lakhani, K. R. 2020. Competing in the age of AI: Strategy and leadership when algorithms run the world. Boston, MA: Harvard Business Review Press. Google ScholarKrakowski, S., Luger, J., & Raisch, S. 2022. Artificial intelligence and the changing sources of competitive advantage. Strategic Management Journal. Forthcoming. Google ScholarLundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S.-I. 2020. From local explanations to global understanding with explainable AI for trees. Nature Machine Intelligence, 2: 56–67. Google ScholarNilsson, N. J. 1998. Artificial intelligence: A new synthesis. San Francisco, CA: Morgan Kaufmann. Google ScholarNiu, S., Liu, Y., Wang, J., & Song, H. 2020. A decade of survey of transfer learning (2010–2020). IEEE Transactions on Artificial Intelligence, 1: 151–166. Google ScholarOstrom, E. 1990. Governing the commons: The evolution of institutions for collective action. Cambridge, U.K.: Cambridge University Press. Google ScholarRathje, J. M., & Katila, R. 2021. Enabling technologies and the role of private firms: A machine learning matching analysis. Strategy Science, 6: 1–109. Google ScholarSenoner, K., Netland, T., & Feuerriegel, S. 2022. Using explainable artificial intelligence to process quality: Evidence from semiconductor manufacturing. Management Science, 68: 5704–5723. Google ScholarShrestha, Y. R., He, V. F., Puranam, P., & von Krogh, G. 2020. Algorithm supported induction for building theory: How can we use prediction models to theorize? Organization Science, 32: 856–880. Google ScholarShrestha, Y. R., Krishna, V., & von Krogh, G. 2021. Augmenting organizational decision-making with deep learning algorithms: Principles, promises, and challenges. Journal of Business Research, 123: 588–603. Google ScholarTang, P. M., Koopman, J., McClean, S. T., Zhang, J. H., Li, C. H., De Cremer, D., Lu, Y., & Ng, C. T. S. 2022. When conscientious employees meet intelligent machines: An integrative approach inspired by complementarity theory and role theory. Academy of Management Journal, 65: 1019–1054. Google ScholarTong, S., Jia, N., Luo, X., & Fang, Z. 2021. The Janus face of artificial intelligence feedback: Deployment versus disclosure effects in employee performance. Strategic Management Journal, 42: 1600–1631. Google ScholarVanneste, B., & Gulati, R. 2022. Generalized trust, external sourcing, and firm performance in economic downturns. Organization Science, 33: 1599–1619. Google ScholarWang, W., Dinh, J. V., Jones, K. S., Upadhyay, S., & Yang, J. 2023. Corporate diversity statements and employees’ online DEI ratings: An unsupervised machine-learning text-mining analysis. Journal of Business and Psychology, 38: 45–61. Google ScholarZhang, C., Xie, Y., Bai, H., Yu, B., Li, W., & Gao, Y. 2021. A survey of federated learning. Knowledge-Based Systems. Google ScholarFiguresReferencesRelatedDetailsCited byPositioning Research on Novel Phenomena: The Winding Road From Periphery to CoreJohn C. Dencker, Marc Gruber, Toyah Miller, Elizabeth D. Rouse and Georg von Krogh17 October 2023 | Academy of Management Journal, Vol. 66, No. 5Publishing Multimethod Research in AMJ: A Review and Best-Practice RecommendationsNed Wellman, Christian Tröster, Matthew Grimes, Quinetta Roberson, Floor Rink and Marc Gruber15 August 2023 | Academy of Management Journal, Vol. 66, No. 4Opening Up AMJ’s Research Methods RepertoireAnn Langley, Emma Bell, Paul Bliese, Curtis LeBaron and Marc Gruber20 June 2023 | Academy of Management Journal, Vol. 66, No. 3 Vol. 66, No. 2 Permissions Metrics in the past 12 months History Published online 19 April 2023 Published in print 1 April 2023 Information© Academy of Management JournalAcknowledgmentsWe are very grateful for comments on earlier versions of this editorial from Savindu Herath, Riitta Katila, Phanish Puranam, and Yash Raj Shrestha. Any errors are our own.Download PDF

Academy of Management JournalVol. 66, No. 6 From the EditorsFree AccessFrom Scarcity to Abundance: Scholars and Scholarship in an Age of Generative Artificial IntelligenceMatthew Grimes, Georg von Krogh, Stefan Feuerriegel, Floor Rink and Marc GruberMatthew GrimesUniversity of Cambridge, Georg von KroghETH Zurich, Stefan FeuerriegelLMU Munich, Floor RinkUniversity of Groningen and Marc GruberEcole Polytechnique Fédérale de LausannePublished Online:19 Dec 2023https://doi.org/10.5465/amj.2023.4006AboutSectionsPDF/EPUB ToolsDownload CitationsAdd to favoritesTrack Citations ShareShare onFacebookTwitterLinkedInRedditEmail Generative artificial intelligence (AI) refers to a class of machine learning technologies that have the capability to generate new content that resembles human-created output, such as images, text, audio, and videos. Within the context of generative AI, much hype has been directed in particular toward large language models, given the associated tools’ ability to generate coherent and contextually relevant text given a user-defined prompt. As editors of a journal devoted to advancing knowledge production in the areas of management and organizations, we believe it is important to recognize the promise of generative AI for advancing such knowledge production. In particular, we see opportunities for scholars to increasingly make use of generative AI to assist with the entire value chain of knowledge production, from synthesis, to creation, to evaluation and translation (Bartunek, Rynes, & Daft, 2001; Kilduff, Mehra, & Dunn, 2011; Van De Ven & Johnson, 2006). Moreover, these same tools promise to increase both the efficiency and rigor of our research methods.Even as we recognize the promise of applying this technology to the craft of scholarly knowledge production, we also recognize the potential perils of doing so. The integrity of the scholarly research process depends upon methodological transparency and reliability. Yet implementation of large language models at scale poses significant risks to such transparency and reliability, even as it introduces additional concerns regarding accuracy, such as “AI hallucinations” and what we are calling “deep research fakes” (Walters & Wilder, 2023). Where the former refers to coherent, contextually relevant, yet false information provided by generative AI that may deceive the researcher, the latter refers to the intentional use of generative AI to create and manipulate data in order to deceive the scholarly community. Moreover, as with all forms of craft, as technology begins to increasingly mediate the human-centered foundations of scholarship, this is likely to raise questions of “authentic” scholarship (Kroezen, Ravasi, Sasaki, Żebrowska, & Suddaby, 2021; Voronov, Foster, Patriotta, & Weber, 2023), and indeed under what conditions human judgment may be necessary to uphold such authenticity.Taken together, the promise and peril of generative AI raise a number of critical questions. First, how might generative AI be used to increase both the quality and quantity of interesting and important scholarship? Second, given the aforementioned tensions surrounding its usage, why should we caution the usage of generative AI? To date, prior editorials and publications from across the academic profession, including a recent Academy of Management Journal editorial (Dwivedi et al., 2023; Susarla, Gopal, Thatcher, & Sarker, 2023; Thorp, 2023; von Krogh, Roberson, & Gruber, 2023), have begun to consider both of these pragmatic and normative questions in light of both the underlying technology and methodology. However, given the broad and nontechnical appeal of generative AI through tools such as ChatGPT and Microsoft Copilot, we see an opportunity to advance this conversation toward a broader and more future-oriented investigation as to the implications of generative AI for the scholarly profession, and for journals like the Academy of Management Journal in particular. To be clear, as we address this third and final question, our intention is not to codify our profession’s response to the future development of this technology, but rather to expose various possible uncertainties that will affect its development and its implications for scholars and scholarship. In other words, we are merely at the beginning of a conversation we expect to be having for many years to come.HOW MIGHT GENERATIVE AI BE USED?In order to understand how generative AI might be used to enhance the quality and quantity of scholarly knowledge production, it is worth deconstructing the scholarly process into its constitutive parts, which we suggest entail (a) knowledge synthesis, (b) knowledge development, (c) knowledge evaluation, and (d) knowledge translation. While some of these steps rely more heavily on “know-what”—the knowledge of data, information, concepts, and theories—other steps rely more on “know-how”—procedural (often tacit) knowledge of how to perform scholarship. In the following paragraphs, we will discuss what we perceive as the promise of generative AI for enhancing each of these steps in the process of scholarship.Knowledge SynthesisThe process of scholarship often starts with the surfacing of theoretically interesting and practically important questions. As such, much of the training provided in PhD programs involves exposure to broad literatures with the aim of familiarizing students with the accumulated body of knowledge in management research, as well as its frontiers, so that they might identify opportunities for meaningfully advancing that knowledge (Dencker, Gruber, Miller, Rouse, & von Krogh, 2023). Such advances are often thought to be more interesting if they are accompanied by theoretical or practical puzzles (Tihanyi, 2020). Moreover, interesting questions are often thought to lie at the intersection of multiple literatures (Grant & Pollock, 2011). While academics often experience efficiencies in knowledge synthesis work within and across literatures over time, such efficiencies can take years to realize. Conversely, emerging generative AI tools now increasingly offer capabilities aimed to increase those efficiencies and the pace at which those efficiencies are realized by scholars (Heidt, 2023). In addition, while the process of discerning or surfacing interesting research questions is often thought to be the purview of human creativity, recent studies have highlighted how GPT-4 can outperform 90% (Haase & Hanel, 2023) and 99% (Guzik, Byrge, & Gilde, 2023) of people on different creative tests. The extent to which generative AI will augment or replace academics in the creative tasks associated with scholarship is a matter of debate (indeed, the authors of this editorial have internally expressed such debate), yet the potential should be taken seriously.Knowledge DevelopmentThe second stage in the process of scholarship often requires knowledge development work, such as an ability to collect, model, and analyze data, and then crafting parsimonious arguments that simply convey complex phenomena and relationships. Here, generative AI promises to profoundly affect the methodological process, which underpins knowledge development work through increases in both efficiency and rigor. With respect to efficiencies, we envision that generative AI will likely lower the barriers to methodological competence, while also speeding up the process of executing on single research tasks. In terms of data collection, we can imagine scholars making use of generative AI to assist with research design as well as instrument design (e.g., survey questionnaires, interview guides). In terms of data analysis, simple prompts or sequences of prompts may be sufficient to engage in sophisticated inductive exploration of large quantitative or qualitative data sets (Gamieldien, Case, & Katz, 2023; Wang et al., 2023). Similarly, simple prompts combined with the above-referenced knowledge synthesis work may be sufficient to not only quickly surface a set of credible yet interesting hypotheses but also follow on to systematically test those hypotheses (von Krogh et al., 2023; Wang et al., 2023). Altogether, these developments enable an accelerated pace of research, leading to what may be called an “instant research paradigm” in management research (Gruber, 2023: 3).With respect to methodological rigor, there are a number of plausible ways in which generative AI might be used to enhance reliability and validity. For instance, for experimental research, generative AI may be used to produce augmented data to test hypotheses in various scenarios, ensuring that findings are consistent and reliable across diverse data sets. It may be used to identify inconsistencies in data, thus ensuring that the data used in analyses is of high quality, and thereby improving the reliability of results. It may be used to create complex simulation models or scenarios that might be difficult to examine in real-world conditions, potentially enhancing external validity. Further, it might be used as a source of researcher feedback, identifying and notifying researchers of potential biases or errors in their methods that could undermine reliability and validity.Beyond the ability to improve data collection and exploration, these tools also offer researchers the potential to improve argumentative clarity while developing hypotheses and propositions. Through outline generation, language enhancement, and proactively identifying and responding to counterarguments, generative AI promises to increase the persuasiveness of scholars’ knowledge development work.Knowledge EvaluationIn the same way that generative AI might enhance the ability of authors to proactively discern deficiencies in both their analysis and argumentation, so too might it assist journal editors and reviewers to identify otherwise overlooked concerns as they engage in knowledge evaluation. As far back as 2006, Chet Miller wrote an editorial for the Academy of Management Journal discussing issues such as reviewer hostility, bias, and dissensus, as well as some of the mitigating approaches journals could continue to take to overcome those issues (Miller, 2006). By providing reviewers and editors with additional feedback on the quality of submitted manuscripts, as well as on the quality of reviews, it is possible that generative AI could be used to further mitigate peer-review concerns. Moreover, such generative AI could be used to match reviewers based on expertise, thus mitigating the likelihood of bias due to lack of topical or methodological familiarity.Knowledge TranslationFinally, the process of scholarship often benefits from knowledge translation work—efforts to ensure the broad accessibility of that knowledge. Within most management journals such translation activities are often focused on converting specific empirical findings into broader theoretical contributions. Yet, increasingly, many scholars have advocated for improvements in such translation efforts aimed at closing the theory–practice divide and ensuring the responsible impact of our scholarship (Ployhart & Bartunek, 2019; Tsui, 2021; Wickert, 2021). Yet, despite these calls, our professional incentives often encourage less focus on such practical translation efforts (Aguinis, Archibold, & Rice, 2022). Additionally, even as the study of management and organizations has globalized, much scholarship remains siloed by language barriers. However, once again, here, generative AI promises to increase the efficiency of such translation efforts through simple prompts (Ahuja et al., 2023; Anderson, Kanneganti, Houk, Holm, & Smith, 2023).WHY SHOULD WE CAUTION USAGE OF GENERATIVE AI?Although there is great promise for using generative AI to increase both the quality and the efficiency of scholarship, the risks are manifold. First, generative AI, and specifically large language models, are trained on massive data sets of text and code. These data can include a variety of sources, such as news articles, social media posts, and academic papers. Using these large data sets, generative AI formulates responses based on patterns rather than specific sources. Yet the actual data and algorithms from which these patterns are deciphered are often “black boxed,” meaning that it is difficult to understand how the conclusions are arrived at. This lack of transparency can make it difficult to assess the reliability and validity of the generated results. Moreover, given this lack of transparency and the constant updating of the model and underlying data structures, it can be difficult for other researchers to replicate results or to identify any potential methodological flaws. This problem is further magnified when humans are interfacing with proprietary systems and are thus removed from the inner workings of the deep learning processes that power generative AI.Second, and relatedly, contemporary versions of generative AI rarely offer and source evidence to substantiate the arguments put forth. Moreover, at least at the moment, generative AI does not (yet) truly understand the use of language in specific contexts in the way humans do (Wittgenstein, 1953). Thus, generative AI would struggle to understand the nuances of a particular academic field or the latest groundbreaking research. Without this deep understanding, it is questionable whether generative AI could reliably provide or source the most current or relevant evidence. In academic research, where precision and accuracy are paramount, this of course raises significant concerns.Third, as previously noted, generative AI has a tendency to hallucinate, whereby it produces persuasive, albeit factually inaccurate and misrepresented, information. This occurs because contemporary generative AI models are probabilistic in nature, such that the AI model makes a prediction about what a meaningful answer might look like yet without applying any forms of logical or contextual reasoning. In academic contexts, such hallucinations could have serious repercussions if they were to go unchecked, including an increase in retractions; damage to author, journal, and professional reputation; as well as ripple effects involved in misleading future research.Fourth, generative AI’s ability to produce human-like text, images, and even videos has opened the door to the creation of “deep fakes” in various contexts, including academic scholarship (Liverpool, 2023). For instance, generative AI could be used to fabricate both quantitative and qualitative data sets that appear legitimate but are entirely fabricated. This could include fake experimental results, surveys, or observational data. Beyond such foundational fabrications, we see the potential of generative AI to produce fake citations, as well as to manipulate images and graphs—all in the service of supporting false claims.Fifth, and finally, we see risks not only with the process of authoring scholarship but also with reviewing and evaluating it. Should reviewers or editors start making increased use of generative AI to summarize, critique, and evaluate manuscripts, we would expect additional challenges for journals in avoiding both “false negatives,” wherein potentially groundbreaking work is rejected, as well as “false positives,” wherein inaccurate, biased, or even fabricated studies are accepted.Taken together, these risks are profound, demanding thoughtful responses from our profession and from the journals that currently uphold the integrity of academic knowledge production. The Academy of Management, as our field’s largest professional organization, will soon publish guidelines regarding the use of generative AI for its set of journals and the conference submissions. At the moment, however, many existing journal policies surrounding generative AI appear to be operating on the assumption that authors, reviewers, and editors will act in good faith. Yet, given both the severity and probability of the aforementioned risks, as well as the highly dynamic nature of the underlying technological developments, we believe such an assumption is inadequate, instead requiring additional and evolving governance designed to deter bad practice. Such governance might, for instance, include AI training for authors and reviewers, specialized review protocols for papers that employ generative AI, the use of AI-assisted verification systems, and periodic audits. While such guardrails would need further debate and discussion, we also see this as an opportunity for further research, whereby management and organizational scholars might investigate both the technological and organizational steps that would reduce the aforementioned risks.WHAT ABOUT THE PROFESSION?In a number of ways, professions are like clubs. Most professions maintain not only barriers to entry but also barriers to advancing within those professions. Such barriers to entry and advancement often entail costly individual efforts to demonstrate relevant expertise and quality output. The academic profession, in this sense, is no different.At present, the academic profession is structured around the presumed scarcity of rigorous scholarly knowledge production, including the generation of new ideas and methods. Journals acquire status within the profession based not only on their impact but also on their exclusivity, as they raise the standards for what qualifies as a novel and important contribution. Tenure-track faculty are competitively hired and promoted based on the perceived quality and quantity of their scholarship, wherein the exclusivity of a given journal is often used as a proxy for assessing that quality. Ultimately, then, the management academic profession is structured around the assumption that scholars have specific knowledge (both “know-what” and “know-how”) that is lacking not only in the public but also among other management professionals, including consultancies.Despite such presumed scarcity, our academic profession is interested in producing and disseminating more knowledge, as long as certain quality standards are upheld. In addition, as noted earlier, generative AI now promises to lower the barriers for individuals to more efficiently engage in rigorous scholarly research. For instance, a recent working paper (Dell’Acqua et al., 2023) highlighted how the use of GPT-4 within Boston Consulting Group not only significantly increased the quality of knowledge outputs but also decreased the gap between the bottom-half and top-half of performers (from 22% to 4%). Whether we expect exponential or more linear rates of improvement in the technology and its associated scholarship-enabling capabilities is, of course, relevant. Yet, in either case, the profession—and indeed our journals—must give thought to the changes on the horizon, the uncertainties that will intersect with those changes, and the internal responses needed. To provoke this consideration we pose two questions, given the potential promise of generative AI to increase both the quantity and quality of scholarship: (a) What does it mean to be a “scholar” when the “know-what” and “know-how” barriers to becoming one are minimized (i.e., anyone who wants to can participate in “scholarship”)? and (b) What does it mean to be a journal that publishes “scholarship” when the field is flooded with manuscripts that meet the highest possible human-mediated standards for (i) practical importance, (ii) theoretical intrigue, and (iii) methodological rigor? We believe that these questions necessitate a degree of scenario planning, in which we attempt to envision and prepare for multiple possible and uncertain futures.Scenario PlanningTo consider different possible scenarios, it is worth clarifying what is it that we assume to be true, and what is it that we assume to be uncertain yet also likely material to understanding the future professional effects of generative AI.First, we believe that it is inevitable that generative AI will become increasingly used to support scholarly knowledge production; we also believe that such usage will simultaneously improve the quality and quantity of knowledge production while also accentuating the risks associated with scholarly malpractice. Second, while there are a host of future uncertainties that are likely to affect the development and adoption of generative AI, we see two particular uncertainties as critical to the technology’s impact on the academic profession: systems transparency and societal regulation.We define systems transparency in this context as the clarity and comprehensibility of the data and methods employed in utilizing generative AI systems, ensuring that other researchers can evaluate, and potentially replicate, the processes and decisions made throughout the research. A recent study, for instance, developed a transparency index by coding 10 foundation models, including many notable generative AI models, on 100 different indicators of transparency related to the resources and data used to build them, the models themselves, and their usage (Bommasani et al., 2023). While scholars will inevitably debate the specific indicators and evidence of such an index, the study usefully highlighted the uncertain and highly varied evolution of generative AI as it pertains to systems transparency. From the perspective of scholarly replicability and reproducibility, the conventional assumption is that we need access to all data and all model details (Shrestha, von Krogh, & Feuerriegel, 2023). Yet, as the study highlighted, most privately owned models fall well short of fulfilling these assumptions (Bommasani et al., 2023).Separately, the second source of uncertainty we explore, societal regulation, can be understood as the extent of governmental legislation restricting the development or usage of generative AI. Many global regulatory bodies, for example, are concerned with the potentially disruptive forces associated with generative AI, and may very well look to cull such disruption. For instance, it may be that societies look to restrict usage to particular domains of society, or to regulate transparency by requiring developers to provide such transparency or performing audits on the algorithms. It may be that the input data for generative AI models become increasingly subject to data privacy restrictions. Finally, it may be that all outputs are regulated, for instance, by way of watermarking.Combined, these two dimensions of uncertainty are likely to implicate our profession of management academia in different ways, which we explore in the following four scenarios (see Figure 1).FIGURE 1 Scenario Illustrations of the Impact of Generative AI on the Academic ProfessionScenario 1—Low systems transparency–low societal regulation.In this scenario, both systems transparency and societal regulation are limited. The underlying algorithms of generative AI systems are proprietary and closely guarded, making it challenging for academics to fully comprehend the system’s functioning. At the same time, there are minimal legislative restrictions on the usage of AI, and it is widely adopted across academia and other sectors without stringent public oversight.The lack of systems transparency poses significant challenges for academia. Researchers may become overly reliant on AI-generated knowledge without fully understanding the underlying processes, leading to potential blind spots and biases in their work (De-Arteaga, Feuerriegel, & Saar-Tsechansky, 2022). Moreover, the absence of societal regulation means that there is limited accountability for the use of AI in research, which could result in the dissemination of misinformation or unethical practices.In a situation of low systems transparency, the credibility and accountability of AI-generated knowledge might be questionable. On the one hand, this could lead to a lack of public trust in AI-generated research; however, it may also lead to an even greater need for academic experts to continue to hold a significant role in knowledge production and verification. The academic profession’s exclusivity may remain relatively stable, as the reliance on human expertise remains crucial to maintaining quality and reliability in academic knowledge.Scenario 2—High systems transparency–low societal regulation.In this scenario, AI developers prioritize systems transparency, allowing for comprehensive understanding of AI systems by experts. This transparency may be underpinned by open-source licensing and communal development principles applied to the creation and deployment of generative AI (Shrestha et al., 2023). However, there is limited societal regulation in place to control the usage of generative AI. The rapid advancement of AI technology outpaces the development of appropriate regulations, and there is a strong push for innovation and unrestricted implementation.With high systems transparency, AI-generated research may still be subject to scrutiny and validation, but the absence of strict societal regulation could lead to widespread use of AI-generated academic content. This scenario might result in an exponential increase in knowledge production, as the process becomes more accessible to a larger number of knowledge workers. Consequently, the academic profession’s exclusivity may decrease as the barriers to knowledge creation and dissemination are significantly lowered.Scenario 3—Low systems transparency–high societal regulation.In this scenario, there is strong societal regulation regarding the usage of generative AI, but the systems themselves remain opaque and difficult to comprehend. Governments and institutions impose strict rules to mitigate the potential risks and negative consequences associated with AI deployment.While the high level of societal regulation aims to protect against misuse and unethical practices, the lack of systems transparency poses challenges for the academic community. Academics may be hesitant to rely on AI-generated knowledge without a deeper understanding of how the systems operate. This cautious approach may limit the integration of AI into research practices, but it also serves as a safeguard against potential pitfalls associated with unscrupulous AI use.In a context of low systems transparency and high societal regulation, human academics may still hold a central role in the academic profession. The focus could shift toward integrating AI-generated knowledge with human expertise, where AI is used to assist human researchers. The exclusivity of the academic profession may be preserved, as the emphasis on maintaining high standards and ethics in knowledge production remains paramount.Scenario 4—High systems transparency–high societal regulation.In this scenario, novel generative AI models are developed with a strong focus on transparency, allowing humans, including academics, to more fully deconstruct and understand the underlying data, models, and usage. The algorithms behind generative AI systems are inspectable, widely shared, well-documented, subject to rigorous scrutiny, and perhaps open-sourced. This level of transparency ensures that biases and errors can be identified and corrected, contributing to high-quality knowledge production.At the same time, there is a high level of societal regulation governing the usage of generative AI. Governments and institutions place strict controls and guidelines on AI research and deployment to prevent misuse and potential harm. Societies are cautious about the rapid advancements of AI and are concerned about its potential impact on social structures, labor markets, and privacy. As a result, AI-generated research and knowledge undergo thorough evaluation and ethical review, protecting academic integrity and the public interest.With high systems transparency, the AI-generated knowledge will be more accountable and more easily scrutinized. This may lead to a situation where the AI-generated research can be effectively challenged and validated by human researchers. Consequently, the role of human academics may shift from traditional knowledge production to knowledge verification, oversight, and translation, and ensuring impact from that knowledge. The exclusivity of the academic profession might decrease as journals are more likely to accept and trust the knowledge production and synthesis from accessible AI tools.CONCLUSIONThe knowledge economy has been responsible for unprecedented growth and human flourishing. Within this knowledge economy, the academic profession has played a pivotal role in the creation of new knowledge, leading to technological advancements, new methodologies, and a deeper understanding of the world around us. Generative AI represents a leap forward both in our ability to engage in such knowledge production and in making the process of knowledge production more accessible to others.Our investigation of the implications of generative AI for management scholarship and for our profession is not meant as a call to arms to defend the profession and its current boundaries. Instead, in the short-term, we view this as a call to prepare ourselves, as well as our current and future PhD students, with the appropriate knowledge not only to use but, more critically, to evaluate algorithmic knowledge production. For instance, scholars need to be trained in the risks of using generative AI such as large language models for scholarship, the ethics of transparent usage, and the methodological competencies for ensuring scholarly integrity while using such powerful yet currently opaque tools. In addition, in the long term, we view this editorial as a call to rethink the distinctive value of our profession in a world of abundant management scholarship. In other words, we suspect that a plausible generative AI-led shift from scarce academic knowledge production to abundant academic knowledge production will inevitably increase the urgency around answering a fundamental question: To what problems in society is management scholarship the (unique) solution?REFERENCESAguinis, H., Archibold, E. E., & Rice, D. B. 2022. Let’s fix our own problem: Quelling the irresponsible research perfect storm. Journal of Management Studies, 59: 1628–1642. Google ScholarAhuja, K., Hada, R., Ochieng, M., Jain, P., Diddee, H., Maina, S., Ganu, T., Segal, S., Axmed, M., & Bali, K. 2023. Mega: Multilingual evaluation of generative AI. arXiv. Google ScholarAnderson, L. B., Kanneganti, D., Houk, M. B., Holm, R. H., & Smith, T. 2023. Generative AI as a tool for environmental health research translation. GeoHealth, 7: e2023GH000875. Google ScholarBartunek, J. M., Rynes, S. L., & Daft, R. L. 2001. Across the great divide: Knowledge creation and transfer between practitioners and academics. Academy of Management Journal, 44: 340–355. Google ScholarBommasani, R., Klyman, K., Longpre, S., Kapoor, S., Maslej, N., Xiong, B., Zhang, D., & Liang, P. 2023. The foundation model transparency index. arXiv. Google ScholarDe-Arteaga, M., Feuerriegel, S., & Saar-Tsechansky, M. 2022. Algorithmic fairness in business analytics: Directions for research and practice. Production and Operations Management, 31: 3749–3770. Google ScholarDell’Acqua, F., McFowland, E., Mollick, E. R., Lifshitz-Assaf, H., Kellogg, K., Rajendran, S., Krayer, L., Candelon, F., & Lakhani, K. 2023. Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality. Working Paper No. 24-013, Harvard Business School Technology & Operations Management Unit. Google ScholarDencker, J. C., Gruber, M., Miller, T., Rouse, E. D., & von Krogh, G. 2023. Positioning research on novel phenomena: The winding road from periphery to core. Academy of Management Journal, 66: 1295–1302.Link , Google ScholarDwivedi, Y. K., Kshetri, N., Hughes, L., Slade, E. L., Jeyaraj, A., Kar, A. K., Baabdullah, A. M., Koohang, A., Raghavan, V., Ahuga, M., Albanna, H., Albashrawi, M. A., Al-Busaidi, A. S., Balakrishnan, J., Barlette, Y., Basu, S., Bose, I., Brooks, L., Buhalis, D., …, Wright, R. 2023. Opinion paper: “So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy. International Journal of Information Management, 71: 102642. Google ScholarGamieldien, Y., Case, J., & Katz, A. 2023. Advancing qualitative analysis: An exploration of the potential of generative AI and NLP in thematic coding. SSRN Electronic Journal. Google ScholarGrant, A. M., & Pollock, T. G. 2011. Publishing in AMJ —Part 3: Setting the Hook. Academy of Management Journal, 54: 873–879.Link , Google ScholarGruber, M. 2023. From the editors—An innovative journal during transformational times: Embarking on the 23rd editorial term. Academy of Management Journal, 66: 1–8.Link , Google ScholarGuzik, E. E., Byrge, C., & Gilde, C. 2023. The originality of machines: AI takes the Torrance Test. Journal of Creativity, 33: 100065. Google ScholarHaase, J., & Hanel, P. H. P. 2023. Artificial muses: Generative artificial intelligence chatbots have risen to human-level creativity. arXiv. Google ScholarHeidt, A. 2023. Artificial-intelligence search engines wrangle academic literature. Nature, 620: 456–457. Google ScholarKilduff, M., Mehra, A., & Dunn, M. B. 2011. From blue sky research to problem solving: A philosophy of science theory of new knowledge production. Academy of Management Review, 36: 297–317.Link , Google ScholarKroezen, J., Ravasi, D., Sasaki, I., Żebrowska, M., & Suddaby, R. 2021. Configurations of craft: Alternative models for organizing work. Academy of Management Annals, 15: 502–536.Link , Google ScholarLiverpool, L. 2023. AI intensifies fight against “paper mills” that churn out fake research. Nature, 618: 222–223. Google ScholarMiller, C. C. 2006. Peer review in the organizational and management sciences: Prevalence and effects of reviewer hostility, bias, and dissensus. Academy of Management Journal, 49: 425–431.Link , Google ScholarPloyhart, R. E., & Bartunek, J. M. 2019. Editors’ comments: There is nothing so theoretical as good practice—A call for phenomenal theory. Academy of Management Review, 44: 493–497.Link , Google ScholarShrestha, Y. R., von Krogh, G., & Feuerriegel, S. 2023. Building open-source AI. Nature Computational Science. Forthcoming. Google ScholarSusarla, A., Gopal, R., Thatcher, J. B., & Sarker, S. 2023. The Janus effect of generative AI: Charting the path for responsible conduct of scholarly activities in information systems. Information Systems Research, 34: 399–408. Google ScholarThorp, H. H. 2023. ChatGPT is fun, but not an author. Science, 379: 313. Google ScholarTihanyi, L. 2020. From “that’s interesting” to “that’s important.” Academy of Management Journal, 63: 329–331.Link , Google ScholarTsui, A. S. 2021. Responsible research and responsible leadership studies. Academy of Management Discoveries, 7: 166–170.Link , Google ScholarVan De Ven, A. H., & Johnson, P. E. 2006. Knowledge for theory and practice. Academy of Management Review, 31: 802–821.Link , Google Scholarvon Krogh, G., Roberson, Q., & Gruber, M. 2023. Recognizing and utilizing novel research opportunities with artificial intelligence. Academy of Management Journal, 66: 367–373.Link , Google ScholarVoronov, M., Foster, W., Patriotta, G., & Weber, K. 2023. Distilling authenticity: Materiality and narratives in Canadian distilleries’ authenticity work. Academy of Management Journal, 66: 1438–1468.Link , Google ScholarWalters, W. H., & Wilder, E. I. 2023. Fabrication and errors in the bibliographic citations generated by ChatGPT. Scientific Reports, 13: 14045. Google ScholarWang, H., Fu, T., Du, Y., Gao, W., Huang, K., Liu, Z., Chandak, P., Liu, S., Van Katwyk, P., Deac, A., Anandkumar, A., Bergen, K., Gomes, C., Ho, S., Kohli, P., Lasnby, J., Leskovec, J., Liu, T., Manrai, A., …, Zitnik, M. 2023. Scientific discovery in the age of artificial intelligence. Nature, 620: 47–60. Google ScholarWickert, C. 2021. Corporate social responsibility research in the Journal of Management Studies: A shift from a business‐centric to a society‐centric focus. Journal of Management Studies, 58: E1–E17. Google ScholarWittgenstein, L. 1953. Philosophical investigations. Chichester, U.K.: Wiley-Blackwell. Google ScholarFiguresReferencesRelatedDetails Vol. 66, No. 6 Permissions Metrics in the past 12 months History Published online 19 December 2023 Published in print 1 December 2023 Information© Academy of Management JournalDownload PDF

Artificial intelligence (AI) offers new possibilities to augment human decision-making under radical uncertainty. This viewpoint commentary explores how AI can relax limits of bounded rationality. It offers a framework for analyzing how AI can support human decision-makers confronted with deep uncertainty by bolstering key decision-making sub-processes. Specifically, AI can help set agendas by scanning environments, formulate problems by providing contextual insights, identify creative alternatives through combinatorial abilities, select options by modelling scenarios and enable rapid experimentation cycles. Connecting the role of AI with the contributions in this special issue, this viewpoint commentary concludes by outlining directions for future research regarding the function of AI and augmented human intelligence in decision-making under conditions of radical uncertainty.