Multi-Agent Interaction and Stability Conditions of Disruptive Innovation by AI Firms in Innovation Ecosystems

Technology enterprises are leveraging artificial intelligence (AI) to foster disruptive innovation, aiming to seize first-mover advantages in technological catch-up and strategic transformation. Most existing studies adopt static research methods such as empirical analysis to explore corporate disruptive innovation from the dimensions of technology, market, organization and value creation. However, few scholars dynamically investigate the impacts of multi-stakeholder interactions on the disruptive innovation of AI enterprises from the perspective of innovation ecosystem by employing evolutionary game theory. Against this backdrop, this paper adopts the evolutionary game approach to explore how the bounded rational strategic interactions among AI enterprises, incumbent enterprises and governments in the innovation ecosystem affect the evolutionary dynamics of AI enterprises’ disruptive innovation behaviors. It also examines under what conditions of benefits, costs, risks and policies the system can evolve toward a stable strategic equilibrium. The findings reveal that the sustainable advancement of disruptive innovation by AI enterprises is not merely driven by the unilateral willingness of individual firms. Instead, it is jointly shaped by the innovation investment of AI enterprises, cooperative responses of incumbent enterprises, and regulatory and supportive policies of governments, as well as comprehensively influenced by base benefits, R&D investment pressure, technology spillover effects and niche competition risks. This research provides theoretical references for improving the innovation governance and policy support system of the AI industry. Future research can further analyze the influence of strategic interactions among more heterogeneous stakeholders on the evolutionary process of disruptive innovation of AI enterprises.