摘要
企业发明者更多地使用生成式人工智能协同创新,将不断更新其在高度创造性任务中的能力认知,这种新型人机协同创新模式的兴起,是否将在根本上影响企业技术创新中的发明者协同?基于社会学习理论,本文探究在生成式人工智能驱动下,技术创新从传统发明者协同模式向新型人机协同模式的演变过程。结果发现:生成式人工智能的引入产生了双重效应,一方面降低发明者对传统社会网络的依赖,使其能够更自主地开展创新活动;另一方面,为处于社会网络边缘的发明者提供更多潜在的创新机会。值得注意的是,这种转变不仅没有降低企业的创新新颖性,反而通过这种新型人机协同激发了更新颖的创新成果,并深刻影响企业间的创新溢出效应。研究结论革新了以往人类专家进行严谨分析、系统梳理与相互交流的创新路径,转而强调一种融合机器自主学习能力的新型人机协同模式。本文的研究不仅有助于理解生成式人工智能在重构企业技术创新进程中的关键作用,也为人工智能时代企业重塑技术创新体系提供相应的启示。
Generative artificial intelligence(AI),a transformative force within the broader AI revolution,is emerging as a critical driver of high-quality enterprise development.Its application is empowering corporate R&D,positioning human-AI collaborative innovation as a cornerstone of innovation process transformation.Yet,this swift practical integration has not been matched by theoretical advances,resulting in a significant knowledge gap and ongoing conceptual controversies.This disparity highlights the necessity of theoretical inquiry into the mechanisms by which generative AI alters innovation paradigms.Consequently,this research holds substantial theoretical and practical value for understanding these impacts and informing the redesign of corporate technological innovation systems.This paper examines,through the lens of social learning theory,the shift from traditional,human-centric inventor collaboration to generative AI-driven human-AI paradigms.It addresses a core research question:Does the advent of generative AI as a collaborative partner fundamentally alter the nature of inventor collaboration within corporate innovation?Traditionally,innovation has relied on inventors leveraging social networks to gather,exchange,and synthesize knowledge.The capacity of generative AI to perform these informational tasks with unprecedented speed and scale,however,challenges the primacy of these traditional interpersonal channels.This study applies social learning theory to analyze this transformation,specifically investigating how the processes of learning and knowledge acquisition are reconfigured when inventors interact with generative AI as a collaborative agent.To reliably estimate the causal effects of generative AI adoption,this study addresses endogeneity by leveraging exogenous shocks within a staggered difference-in-differences design,strengthened by Propensity Score Matching.We construct proxy variables for inventors’“accessibility”and“exposure”to generative AI collaboration opportunities.Guided by social learning theory,which explains learning through interaction,we analyze changes in knowledge acquisition,information sharing,and behavioral patterns across innovation ecosystems.The empirical analysis uses detailed data on corporate inventors and their outputs,linked to generative AI adoption events,and employs regression models with firm-level and quarterly fixed effects.The main findings reveal a dual effect of generative AI adoption on collaboration networks:It reduces inventors’reliance on traditional social networks for information,enabling more autonomous innovation,while simultaneously providing increased potential innovation opportunities for inventors previously on the periphery of social networks.Contrary to potential concerns,the shift towards human-AI collaboration does not diminish innovation novelty but instead stimulates the generation of more novel technological outcomes.Furthermore,generative AI acts as a bridge for knowledge integration,facilitating the flow of novel knowledge across boundaries and significantly enhancing innovation spillover effects compared to traditional models.Policy recommendations suggest that enterprises should actively integrate generative AI as a core collaborative partner into their innovation systems,strategically leveraging its capabilities for knowledge integration and open innovation.Policymakers are advised to formulate policies that encourage the responsible development and adoption of generative AI in R&D,support the creation of an inclusive innovation ecosystem,and improve intellectual property frameworks.The key innovations and contributions of this research include its theoretical contribution to micro-level AI-innovation research by introducing a social learning theory framework,providing novel insights into how generative AI affects social network dynamics within innovation,particularly in specific social contexts,and enhancing the understanding of innovation spillovers in human-AI collaboration by demonstrating how generative AI facilitates knowledge diffusion.In conclusion,this research demonstrates that generative AI is fundamentally reshaping corporate technological innovation by transforming collaborative patterns,with significant implications for novelty,network dynamics,and knowledge diffusion.
作者
王向前
WANG Xiangqian(Business School,Nankai University)
出处
《经济研究》
北大核心
2025年第9期156-176,共21页
Economic Research Journal
关键词
生成式人工智能
技术创新
社会学习理论
发明者协同
Generative Artificial Intelligence
Technological Innovation
Social Learning Theory
Inventor Collaboration
