An Engineering Definition of AGI: Complex-Project Closed-Loop Capability as an Assessment Criterion

Artificial General Intelligence (AGI) has long lacked a stable and operational definition. Existing definitions variously emphasize human-like behavior, subjective mental states, broad cognitive capability, human-level performance, economic value, or social impact, yet they still leave an engineering question unresolved: when does AI transition from a high-capability tool to a system capable of independently assuming responsibility for complex productive work? This paper shifts the unit of assessment from local capabilities and individual tasks to complex-project closed-loop capability, and proposes a working definition of engineering AGI: given an objective, available resources, safety boundaries, and initial context, an AI system can, without continuous human steering, reliably complete most typical complex projects that would otherwise require sustained collaboration by ordinary human teams, while producing verifiable deliverables. This definition does not attempt to resolve philosophical questions such as consciousness or subjective experience. Instead, it provides a system-level assessment framework for AI engineering, knowledge-work automation, and productivity assessment that supports further operationalization, cross-system comparison, and progressive approximation.