Unintended consequences in Information Systems (IS) research are typically examined only after they have surfaced in practice, often treated as isolated instances rather than being systematically identified and understood. This reactive stance constrains our ability to anticipate risks, recognize emerging opportunities, and build cumulative insights across studies. To assess the current state of IS research on unintended consequences, we conducted a scoping review of literature in the Association for Information Systems' Senior Scholars' Basket of Journals. Our review highlights significant gaps in how unintended consequences are defined, theorized, and systematically compared. To address these gaps, we propose three foundational considerations: (1) conceptual clarity, to refine definitions and delineate boundaries; (2) relational configurations, to analyze the interplay between actors, technologies, and contexts; and (3) temporal configurations, to capture the evolving nature of consequences over time. Building on these dimensions, we introduce a set of guiding questions designed to provide shared analytical vocabulary rather than a prescriptive framework. These questions enable researchers to more systematically identify, categorize, and compare unintended consequences across contexts, thereby fostering theoretical precision, facilitating cross-contextual learning, and supporting a more anticipatory and comprehensive understanding of the phenomenon.

Large Language Models (LLMs) are increasingly embedded in organisational workflows, serving as decision-making tools and proxies for human behaviour as silicon samples. While these models offer significant potential, emerging research highlights concerns about biases in LLM-generated outputs, raising questions about their reliability in complex decision-making contexts. To explore how LLMs respond to challenges in Information Systems (IS) scenarios, we examine ChatGPT’s decision-making in three experimental tasks from the IS literature: identifying phishing threats, making product launch decisions, and managing IT projects. Crucially, we test the impact of role assignment, a prompt engineering technique, on guiding ChatGPT towards behavioural or rational decision approaches. Our findings reveal that ChatGPT often behaves like human decision-makers when prompted to assume a human role, demonstrating susceptibility to similar biases. However, when instructed to act like AI, ChatGPT exhibited greater consistency and reduced susceptibility to behavioural factors. These results suggest that subtle prompt variations can significantly influence decision-making outcomes. This study contributes to the growing literature on LLMs by demonstrating their dual potential to mirror human behaviour and improve decision-making reliability in IS contexts, highlighting how LLMs can enhance efficiency and reliability in organisational decision-making.

The start-up ecosystem in Australasia, encompassing Australia and New Zealand, has demonstrated significant potential for innovation and growth. However, start-ups in this region face substantial challenges in achieving sustained growth, including limited access to capital, insufficient market knowledge, and the complexities of navigating regulatory environments. This article explores these unique challenges and examines how Information Technology (IT) service providers have emerged as central players in the start-up ecosystem, offering strategic consulting services that encompass growth and expansion. By analysing the case of Tata Consultancy Services' (TCS) Business as a Service (BaaS), the article investigates the dynamics of power and the strategies utilized by both start-ups and service providers to sustain beneficial positions within these partnerships. Utilizing a multi-theoretical lens, the article examines how strategic partnerships between start-ups and IT service providers can foster a thriving entrepreneurial ecosystem in Australasia. The findings emphasize the importance of these partnerships in creating mutual dependencies that enable growth and expansion. However, it also highlights the potential risks associated with such dependencies, urging start-ups to carefully weigh the benefits and challenges of engaging with IT service providers for strategic consulting services. This controversy accentuates the critical debate on whether start-ups should embrace these partnerships or avoid them due to the inherent risks, turning it into a key decision point for their strategic decisions.

Process automation has been a cornerstone of organisational success, driving competitiveness through enhanced effectiveness and efficiency. While automation remains a key focus, the rise of artificial intelligence (AI), machine learning (ML) and the Internet of Things (IoT) provides entirely new process design options and as such demands new skills and approaches in Business Process Management (BPM) to maintain relevance and competitiveness. This paper extends current BPM paradigms by introducing the concept of “process autonomisation,” a new paradigm that empowers organisations to make decisions autonomously using real-time data and adaptive business contexts. By critically examining existing BPM methodologies and highlighting key innovations, we propose strategies for evolving process management frameworks - emphasising agility, data-driven decision-making, and seamless digital integration. To thrive in the digital age, we argue, organisations must rethink their BPM practices. We call on both academic and industry stakeholders to collaborate on pioneering advancements in BPM and co-develop a roadmap that will help organisations align their process management with the ongoing digital transformation reshaping industries.

Robotic process automation (RPA) is increasingly adopted as a relatively inexpensive automation solution to reduce routinised and repetitive tasks and to initiate an organisation’s broader automation programme. Prior research has focused on highlighting RPA benefits for organisations with suggestions on how to maximise benefits and avoid challenges in RPA implementation. There is less understanding of the emergent and dynamic nature of RPA implementation. Drawing on key elements of socio-technical change, we conducted a process study of RPA implementation in a university. From our analysis, we identified five process patterns: initiation, mobilisation, configuration, adaptation, and evaluation, each of which has different implications for organisational trajectories of RPA implementation. Our findings also offer insights into how the changing role of RPA as an epistemic, technical, and agentic object is intertwined with the dynamics of automation and augmentation in RPA’s conception, development, incorporation into work routines, and evaluation of the initiative’s future.

'HyFlex' is a catchword for course designs in higher education that provide students with the opportunity to attend course sessions in-person as well as remotely ('hybrid') and to change the mode of attending every week based on their circumstances and preferences ('flexible'). Due to the need to teach the complex skill of conceptual enterprise modelling (CEM) through practical exercises and feedback, HyFlex course designs for CEM courses pose several challenges to typical CEM-related learning outcomes. This paper proposes a set of design requirements, principles, and features for a comprehensive HyFlex course architecture for CEM courses. The design requirements, principles, and features evolved through and are evaluated against several iterations of CEM courses across two different programmes (undergraduate and postgraduate), student numbers (ranging from 17 to 121), settings (lecture & workshops, block mode, etc.), and course topics (business process and enterprise architecture modelling). The student performances and course evaluations indicate that the offered courses that instantiated the presented course architecture have been both effective and appreciated by the students. Other instructors can draw on our course architecture to design or adapt their own CEM courses – but also, to an extent, courses with other topics – to the HyFlex paradigm.

Managing cyber risk in the supply chain represents one of the most significant challenges in cyber risk management. The paper studies how organizations assess supplier cyber risk. We used a mixed-method approach. We conducted 33 semi-structured interviews with cybersecurity experts from various organizations closely involved in supplier cyber risk assessments, as well as consultants. We complemented our qualitative findings by surveying 53 security experts about their supplier cyber risk assessment. Based on the qualitative findings, we formulate a process theory of supplier cyber risk assessment. This theory explains how organizations assess supplier cyber risk and which contextual factors affect the maturity of cyber risk assessment and monitoring. The quantitative analysis supports the qualitative findings and suggests that the process can effectively identify risky suppliers. The paper sheds light on challenges and strategies associated with supply chain cyber risk assessment. The practical implications of our findings offer actionable insights for organizations seeking to enhance their cyber supply chain risk management.

Imagine a world where the boundaries between physical and digital realities dissolve, creating immersive experiences that transform how we interact. This potential lies at the heart of the Metaverse, and a confluence of technologies such as AR, VR, AI and blockchain will be the key to unlocking it. We employ a novel combination of horizon scanning and narrative development to explore the transformative role of AR within the Metaverse. This approach reveals potential technology-driven futures, highlighting emerging trends and disruptions not readily apparent through traditional forecasting methods. Our narratives of the future offer surprising glimpses into potential sociotechnical futures, informing a research agenda for both industry and academia.

Humanity is linguistically diverse, but science is not. Academic success requires English-language mastery. Every major conference, every major journal – even this one – assumes it. English-language bias in science is so strong that it is taken for granted by most scientists and scientific associations, never talked about nor addressed. This is unfair and creates great costs and missed opportunities. It is also unnecessary. Artificial intelligence (AI) translation tools are becoming very good, very fast, allowing us to foresee a multilingual science. Our provocation to readers is: How should we harness AI translation tools for a more impactful, inclusive science? This is a challenge ideally suited to Information Systems scholars because it involves designing sociotechnical artifacts and practices for a better future. To demonstrate feasibility, this article went through a multilingual review process and is published in five languages, all enabled by AI translation.