Rising disaster risks associated with climate change, environmental change, and rapid urbanization have heightened the need for innovative technologies that strengthen disaster preparedness and resilience. Among these technologies, artificial intelligence (AI) has attracted growing attention for its ability to improve risk prediction, early warning, and emergency response. This study examines the effect of AI technology innovation on disaster risk across 24 countries from 2002 to 2023. Using the bias-corrected linear dynamic panel data estimator, the results show that AI technology innovation significantly reduces disaster risk. Financial development strengthens the disaster-reducing effect of AI, whereas globalization does not significantly moderate this relationship. Although AI technological innovation contributes positively to institutional quality, the mediation results indicate that institutional quality is not a significant transmission mechanism through which AI affects disaster risk. The heterogeneity analysis further reveals that the disaster-reducing effect of AI technology innovation is stronger and significant in developing and low-disaster-exposure countries. The findings underscore the importance of sustained investment in AI innovation, complementary financial systems that facilitate technology deployment, and stronger institutional and technical capacity to translate AI innovation into effective disaster risk reduction strategies.
Responsible innovation represents a new innovation paradigm that balances profit growth, social values, and ethical norms, playing a supportive role in advancing sustainable development. As the fundamental business units of innovation, teams can rapidly respond to corporate strategies to meet stakeholder expectations. However, research on how to incentivize teams to engage in responsible innovation remains scarce. Recognizing the institutional power of socially responsible human resource management (SRHRM) for teams, we draw on input-process-outcome theory to investigate whether, how, and when SRHRM shapes team responsible innovation. Using an independent sample of 648 team leaders from 5 technology-intensive firms, we validate the team responsible innovation scale and distinguish it from related constructs. In a separate multilevel field study involving 55 human resource managers, 213 team leaders, and 1145 team members, the results confirm that SRHRM is positively associated with team responsible innovation and that this relationship is mediated by team reflexivity and stakeholder learning. Furthermore, we find that responsible leadership strengthens the indirect effects of SRHRM on team responsible innovation via team reflexivity and stakeholder learning. Theoretical and practical implications are discussed.
Smart grid technologies such as renewable energy power forecasting evolve continuously within the renewable energy market. Accurately forecasting their strategic value is essential for technology marketization planning. However, technologies and market are mutually constitutive and co-evolving—a characteristic we term strategic salience, defined as the strategic value a technology exhibits through both its intrinsic importance and relational strength within the system. Technologies with high strategic salience typically attract more resources, accelerating their own evolution while reshaping the strategic value of other system components, thereby introducing substantial complexity and uncertainty into technology marketization planning. Therefore, accurately characterizing and forecasting the strategic salience of market and technologies is a challenge. To address this, we propose a two-stage framework for mapping and forecasting the strategic salience. First, we utilize Large Language Model (LLM) with Retrieval-Augmented Generation (RAG) to construct a demand-technology dynamic graph, from which we extract features to quantitatively describe strategic salience. Second, we design a GNN model that jointly learns interactions between node and edge features, framing the task as a joint forecasting problem to accurately forecast strategic salience trends. Experimental results based on 48,546 smart grid technology documents demonstrate that our framework significantly outperforms benchmark models in forecasting the strategic salience.
This paper investigates the ESG commitment of early-stage firms. Although current literature suggests that the poor ESG commitment of these firms is attributable to their limited resources, using a life-cycle-based perspective in analyzing ESG engagement, we suggest that the strategic and organizational characteristics of early-stage firms impact managerial priorities regardless of resource availability. Based on a sample of companies over the period 2013–2022, our results demonstrate that the ESG commitment of highly profitable early-stage firms is lower than that of highly profitable late-stage firms. Furthermore, the ESG commitment of highly profitable early-stage firms is not different from that of their low-profitable counterparts. Our findings challenge one of the most established explanations in the ESG literature, showing that alleviating financial constraints of early-stage firms does not automatically lead to greater ESG commitment. The results are robust to several controls. The implications of the work are also discussed.
As AI becomes increasingly embedded in supply chain operations, understanding how supply chain managers engage with these technologies in a responsible and context-sensitive manner is essential for advancing sustainable supply chain. Existing research tends to treat AI use as a trait-based phenomenon, overlooking the behavioral dynamics the same manager could exhibit across varied situations. This perspective article introduces the concept of the responsible AI use signature (RAI Signature), grounded in Cognitive-Affective Personality System theory (CAPS), to capture the stable yet dynamic patterns of responsible AI behavior demonstrated by the same manager across task, relational, and geographical contexts. A multi-level conceptual framework is developed to explain how RAI signature emerges and how it influences sustainable supply chain outcomes. Drawing on CAPS, this perspective article argues that RAI signature reflects the activation of cognitive-affective units in response to situational cues, moderated by individual, organizational, and technological factors. Three main propositions are developed to advance theory by shifting the focus from static competencies to within-person behavioral variability, offering a new lens for understanding, measuring, and developing managerial capacity in AI-enabled supply chains. This article contributes to research at the intersection of responsible AI, supply chain sustainability, and human-centered digital transformation.
