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

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

Artificial general intelligence (AGI) has long lacked a stable and operational definition. Existing definitionsvariously emphasize human-like behavior, subjective mind, broad cognitive ability, human-levelperformance, economic value, or social impact, yet they still struggle to answer an engineering question:when does AI cease to be a high-capability tool and become a system able to assume independentresponsibility for complex production? This paper shifts the unit of judgment from local capabilities andindividual tasks to complex-project closure and proposes a working definition of engineering AGI: givengoals, resources, safety boundaries, and initial context, and where acceptance boundaries can be establishedindependently, an AI system can, without continuous human direction, reliably complete most representativecomplex projects that previously required long-term collaboration by ordinary human teams, whileproducing verifiable deliverables. The definition primarily covers complex projects whose solution paths maybe unknown but that can be closed by external goals and evaluation conditions; it does not require thesystem to consistently produce original contributions beyond the human frontier on open problems that lacka stable formulation or evaluation standard. The paper's structural-increment claim is that the key unit forjudging AGI should move from local capability to continuity of responsibility in closable complex projects,and should be clearly distinguished from open-frontier innovation. Its directional-value claim is that thisboundary makes AGI continuously measurable through project coverage, human intervention, deliveryreliability, and independent acceptance evidence, providing a comparable, verifiable, and incrementallyapproachable system-level target for AI engineering, knowledge-work automation, and productivityassessment. The definition does not attempt to resolve philosophical questions such as consciousness orsubjective experience.