The Rebel Minds Human Layer Framework on AI Adoption (RMHLF): Psychological Safety, Self-Efficacy, and the Gap Between Adoption and Return

Organizations have adopted generative artificial intelligence at scale while comparatively few report measurable financial return. Stanford's AI Index places organizational adoption at 88%; McKinsey's State of AI survey finds only 39% of organizations can attribute any bottom-line impact to it, and roughly 6% capture significant value. Prevailing explanations target the technology — model quality, integration, prompt design. This paper argues the binding constraint is elsewhere: in the human layer, defined here as the set of beliefs, conditions, and capabilities that determine whether people take up an AI tool and derive return from it. The framework advances a four-link chain: psychological safety → adoption → self-efficacy → return. Its first link rests on published evidence. In a study of 2,257 employees at a global consulting firm, each one-unit increase in psychological safety raised the odds of AI adoption by approximately 30%, an effect stable across seniority, experience, and region. Critically, the same work found psychological safety did not predict depth or duration of subsequent use. Adoption is therefore necessary but not sufficient for return. That asymmetry defines what this paper terms the sufficiency gap: the interval between a tool being adopted and the adoption producing value. The framework's central proposition is that this gap is substantially a self-efficacy gap — a matter of task-specific capability belief, which decades of research establish as both measurable and buildable, and whose antecedents are addressable through training design. This proposition is advanced as a testable claim grounded in established self-efficacy and training-transfer research, not as a settled empirical finding, and the boundary between the two is marked throughout. Keywords: AI adoption; artificial intelligence; psychological safety; self-efficacy; technology adoption; training transfer; industrial-organizational psychology; organizational behavior; human factors; technology acceptance.