Building Meta-Dynamic Capabilities Through AI-HI Collaboration: Experimental Evidence from Multinational Organizations in Disaster Response Operations

The rise in large language models (LLMs) has sparked renewed interest in how firms, particularly multinational aid organizations, can enhance learning related to meta-dynamic capabilities (DCs), such as agility, sensing, and adaptation, in response to disasters and humanitarian crises. A key strategic priority is developing meta-rules that combine general engagement frameworks with locally tailored action plans, grounded in cultural and institutional contexts. LLMs offer potential in supporting this need, but premature deployment risks harmful or misleading outcomes. This underscores the critical importance of collaboration between artificial and human intelligence (AI-HI). While AI brings computational power, it lacks the tacit knowledge—encompassing cultural, contextual, and intuitive understanding—that is essential in high-stakes, unpredictable environments. Our experimental study provides two core insights: (1) AI alone cannot effectively handle tasks requiring tacit knowledge, and (2) AI-HI collaboration thrives when human input guides AI using deep awareness of local social and political dynamics. We contribute to the discourse on dynamic capabilities in multinational contexts during catastrophic situations by offering practical strategies to support successful AI-HI partnerships and a framework for organizations aiming to enhance meta-DCs through responsible, human-centered use of disruptive technologies. Our findings clarify how the international dimensions of these capabilities influence their effectiveness across diverse cultural and institutional environments.