Contenu du sommaire : Artificial Intelligence and Sustainability

The aim of this article is twofold: 1. Suggest an overview of current knowledge and understanding of both concepts and 2. Present the six contributions and their positioning in relation to current the literature linked to artificial intelligence and sustainability. For that, we use different but complementary sources. First, we ask artificial intelligence to reveal the mainstream view. Then we call on human intelligence to provide a critical perspective. Finally, we carry out a bibliometric analysis using the SCOPUS database and two different statistical analyses (the CCA – co-citation analysis, the BCA – bibliographic coupling analysis). The diversity of the sources used, and their complementarity allow us to propose a holistic vision of the subject, highlighting the concerns that surround it and identifying future avenues of research for academics. The articles selected in this special issue fill some of the gaps raised and call for further research.JEL Codes: O33, O44, Q55, Q56

This paper extends disruptive innovation theory by employing bibliometric analysis to study the scope, magnitude, and interconnections of multilevel socio-technical disruptions termed “systemic disruption”. Focusing on the context of the space ecosystem from 2005 to 2021, characterized by many disruptions known as New Space, our analysis reveals six interconnected disruption clusters: new technology, new policy, new market, new entrant, new process, and new funding approach. The most influential disruptions are new technology, new policy, and new market. Notably, interconnections between new technology and new policy disruptions shape the scope and magnitude of systemic disruption through a co-evolution mechanism. This study advances disruptive innovation theory beyond technology disruptions and the level of individual organizations, expanding its applicability to systemic contexts. Furthermore, it provides guidance to researchers on applying bibliometric analysis in the field of innovation, while practitioners can leverage the research to plan innovation agendas considering the systemic nature of disruptions. JEL CODES: O33, L90, B0

This study examines the structural changes in the German automotive research network to gain deeper insights into network transformation during regime change. Using a separable temporal exponential random graph model (STERGM), the author explores the influence of various factors on link formation and dissolution, including actor characteristics and general structural measures. Analyzing networking behavior over different time periods sheds light on the changing significance of these factors, particularly during the technological regime change in the German automotive sector. Findings reveal that link formation during regime change exhibits an increasing tendency toward cognitive and institutional distance between actors, with shorter durations of connections. Additionally, organizational size, urban location, and geographical proximity positively impact link formation. JEL CODES: O30, O33, L14

The conditions under which R&D resources are allocated either to individual or collective R&D projects are largely unexplored. We contribute to closing this gap by asking under which conditions firms – each of which occupies a unique position along the automotive industry value chain – may overcome innovation barriers and spend scarce resources for collective R&D projects. We use a framed laboratory experiment to scrutinize the influence of different situations on the decision to spend the R&D budget for individual or collective R&D projects. The framing originates from a real-world case study of the massive metal forging industry. We identify constellations that support budget spending for collaborative purposes, e.g., sequential decision-making, which also increase the overall welfare, even in the case of unequally distributed R&D budgets.JEL Codes: O31, O32, C91.

Cognitive biases within Artificial Intelligence are a common phenomenon. Before this background, we asked ChatGPT what the cognitive biases of AI are. In this article, we critically reflect on the response we got from Generative AI, and discuss cognitive biases in general. We then focus on concrete implications for handling cognitive biases. Such as Algorithms and techniques that can help AI systems identify and reduce bias could be implemented, and standardized techniques to check AI systems for bias before implementation. We also highlight the importance of realizing procedures for ongoing observation and input to identify and rectify prejudices in AI systems once they are put into use. We conclude that solving such biases calls for an all-encompassing strategy that incorporates experts from other fields and places an emphasis on moral issues in addition to technological fixes. By doing this, we can encourage the creation of AI technologies that are not just advanced but also ethical, equitable, and beneficial to society as a whole.