Reclaiming Human Innovative Intelligence and Agency in the Age of Artificial Intelligence

The rapid expansion of artificial intelligence (AI) has changed the economics of information retrieval, synthesis and idea generation. It has not, however, eliminated the need to decide which problems merit attention, which assumptions constrain inquiry, which conclusions are justified by evidence and which futures should be pursued. This paper argues that human innovative capability should be deliberately developed alongside AI deployment. It introduces the Meta-Box as a heuristic for examining nested assumptions at the levels of task, process, system, purpose and underlying human need. It then proposes a Human Innovation Lab: a reproducible, evidence-informed training environment combining disciplined curiosity, assumption interrogation, analogical transfer, radical problem reframing, intelligent dissent, structured human–AI cooperation and experimental verification. The proposal is critically situated against Creative Problem Solving, design thinking, TRIZ, Synectics and related methods. The published creativity-training literature provides qualified support for trainability, but methodological weaknesses and limited evidence of long-term workplace transfer preclude strong efficacy claims. Accordingly, a falsifiable research program is specified, including blinded assessment, active comparators, unfamiliar transfer tasks and longitudinal workplace outcomes. The paper advances a normative and empirical proposition: the value of AI will be greater when organizations cultivate people capable not merely of obtaining answers but of discovering, challenging and reconstructing the questions that govern action.