Academic Summary: This article draws on the attention‐based view (ABV) to examine whether, how, and under what conditions top management team (TMT) attention to responsible artificial intelligence (AI) influences firm innovation. We developed a 480‐word responsible AI dictionary grounded in 155 academic sources and 527 corporate case descriptions, and applied it to 2452 S&P 500 earnings call transcripts (2011–2021) using natural language processing (NLP) and large language model (LLM) techniques, yielding 2670 firm‐year observations. Linking these measures to US patent data, we find that greater responsible AI attention predicts more and higher‐impact patents. The effect is stronger in low‐technology industries and under short‐term investor pressure, while the presence of a chief technology officer (CTO) does not amplify it. Mechanism analyses reveal that responsible AI attention fosters innovation by increasing investment in AI‐relevant human capital and mitigating innovation risk. Theoretically, this article enriches the AI and innovation management literature by positioning responsible AI attention as a dynamic strategic asset that mobilizes resources, reduces risk, and enables contextual adaptation. Practically, findings suggest that firms can strengthen innovation by prioritizing managerial attention to responsible AI, distributing responsibility beyond technical specialists, balancing ethical safeguards with strategic flexibility, and aligning governance with investor and industry conditions. Managerial Summary: This article examines how managerial attention to responsible artificial intelligence (AI) can enhance firm innovation. Using text analytics on 2452 earnings call transcripts from S&P 500 firms (2011–2021) and a panel of 2670 firm‐year observations linked to patent outcomes, we show that firms whose top management teams (TMT) devote greater attention to responsible AI produce more and higher‐impact patents. This effect is stronger in low‐technology industries and when firms face short‐term investor pressure; it is not amplified by having a chief technology officer (CTO). In practice, sustained attention to responsible AI tends to build AI‐related skills and reduce project risk, thereby supporting a more reliable innovation pipeline. Executives should treat responsible AI as a strategic priority rather than a compliance task by establishing cross‐functional governance, investing in role‐based governance training, and sharing accountability across the C‐suite. Innovation managers can embed ethics checkpoints (bias audits, design reviews) into project workflows to enhance stability and organizational learning. Policymakers can reinforce responsible innovation by providing clear regulatory frameworks and incentives that align ethical safeguards with competitiveness. Together, these actions can help build more durable organizational capability for responsible innovation and support long‐term performance and adaptation to ongoing technological change.

Currently, on the Chinese internet, the term “involution” refers to intense but meaningless competition for limited resources, also known as “the inflation of effort.” Involution is a prevalent phenomenon in various social competition scenarios and can lead to significant social contradictions. On the other hand“, lying flat” refers to a conscious choice by individuals to withdraw from intense competition and pursue a self-centered and hollow existence. The propagation of lying flat culture may lead to nihilism or irrationalism. However, existing literature on involution and lying flat has two main limitations. First, previous studies have separately examined involution and lying flat as two opposite yet closely related outcomes of intense social competition. This approach fails to provide a comprehensive and in-depth understanding of involution and lying flat in the context of increasingly fierce social competition. Second, existing literature predominantly adopts value-based payoff allocation methods, disregarding the prevalence of rank-based payoff allocation methods. Consequently, the other prevalent rank-based allocation (RA) method remains understudied. To address these limitations, we propose an evolutionary game model on a square lattice to abstract the fierce social competition between involution, cooperation, and lying flat. Our experimental results demonstrate that various factors, including social resources, the cost of involution, the relative utility of the involution strategy, and the variance of agents’ competition abilities, play distinct yet significant roles in social competition. Furthermore, these factors exhibit interaction effects on the levels of cooperation, involution, and lying flat. The rank threshold ( $T)$ positively affects the level of cooperation, exhibits a U-shaped effect on the level of lying flat, and an inverted U-shaped effect on the level of involution. The RA rule tends to result in a higher level of involution, whereas the value-based rule tends to result in a higher level of lying flat. These findings remain robust across different network sizes. Our study provides insights into strategies for suppressing the levels of involution and lying flat in social competitions.

Purpose: This study investigates the multifaceted impact of generative artificial intelligence (GenAI) usage on scholars' mental health. It examines the mediating role of proactive innovation behavior (PIB) and the moderating impact of artificial intelligence (AI) literacy, focusing on how AI usage intensity and diversity relate to anxiety, stress and depression. Design/methodology/approach: The study analyzes survey data collected from 586 Chinese academic scholars. Structural Equation Modeling methods were employed to test the direct relationships between AI usage dimensions and mental health outcomes, the mediating effect of PIB and the moderating effect of AI literacy on these relationships. Findings: AI usage intensity is negatively associated with anxiety and stress, while greater AI tools diversity predicts lower depression. Both usage intensity and diversity foster PIB, which mediates the relationship with reduced anxiety, stress and depression. AI literacy moderates these links, strengthening the direct mental health benefits of intensity but weakening the intensity-PIB link and the adverse effects of diversity on anxiety/depression. Originality/value: The study highlights the complex interplay between AI engagement dimensions, scholars' innovative behaviors and AI proficiency concerning mental health. It contributes a nuanced understanding to the technology and mental health literature regarding AI integration in academia, offering theoretical insights into innovation's mediating role and digital literacy's conditional effect.

The high-quality and sustainable development of the digital economy cannot be achieved without the support of the policy system. The purpose of this article is to scientifically analyze the laws and characteristics of the policy system that supports the sustainability of the digital economy. Specifically, based on public policy theory, this article constructs an analytical framework of “policy objectives—policy tools—organizational management” and incorporates the “central-local” relationship with Chinese characteristics into the framework. Meanwhile, text analysis was used to quantitatively analyze 971 digital economy policies issued by the central and local governments in China from 2000 to 2022. Research shows that the central government of China has taken the lead in introducing digital economy policies, and local governments are able to respond quickly. The central government’s digital economy policy goals fluctuate and change, while local governments pay more stable attention to the digital product service industry. Both central and local governments widely use supply oriented policy tools and environmental-oriented policy tools, while demand-oriented policy tools are not widely applied. The vertical relationship between central and local data management agencies is still unclear, and the degree of cross departmental and cross regional cooperation is not high. Meanwhile, empirical analysis based on the two-way fixed effects model shows that technological innovation, R&D funding, and industrial structure have a significant impact on changes in digital economy policies. This article summarizes the characteristics of the evolution of digital economy policies in the Chinese context, providing policy design references for promoting the sustainability of the digital economy and methodological insights for policy research in other fields.

Involution is an inefficient equilibrium of positional competition in which intensified competitive input fails to generate proportional positional advancement under scarce social resources and weak exit options. This study conceptualizes involution as non-advancement-generating input under positional scarcity and develops a hybrid model combining rule-based agent-based modeling (ABM) with large language model (LLM) assisted smart agent-based modeling (SABM). Agents in SABM make bounded movement decisions and provide natural-language rationales. ABM enables scalable pattern assessment and sensitivity analysis. The results identify five patterns: homogeneous crowding and the embedded-catfish agent intensify involution, whereas agent-attribute heterogeneity, advancement-path diversification, and competitive-exit activation mitigate it. The catfish effect is conditional rather than linear: a single sufficiently embedded catfish agent can amplify involution, while a sufficiently larger number of catfish agents may shift the system toward broader agent-attribute heterogeneity and mitigate involution. These findings clarify how positional scarcity, homogeneity, embeddedness, catfish-agent quantity, path structure, and exit options jointly shape involution dynamics.