Adaptive Labor Infrastructure for AI Disruption: The Hybrid Agora: collective meta-intelligence, human-agent teams and a virtuous learning circle between communities and machines

Abstract

Artificial intelligence already performs much of the repetitive cognitive work in closed domains faster and at lower cost. The policy and research conversation now asks how workers and AI build expertise together, whether reskilling and reemployment can scale at the pace of disruption, and what public and private investment in workforce transitions actually delivers. This paper addresses a question those agendas leave open: where, with whom and for what purpose a person whose work is changing retrains, and what the human community and the machine can learn from each other in the process. It proposes an Adaptive Labor Infrastructure for AI disruption (ALI): a territorial system whose unit is the Collective Meta-intelligence Center (CMC), conceived as a hybrid agora — the oral, dialogic and peripatetic space of the classical city, updated with AI agents as team members and with measured outcomes. The paper makes five nameable contributions. (1) It reformulates Sternberg et al.’s (2021) individual meta-intelligence as a group property with two axes, capacity and orientation, whose product, vital lucidity, is measurable. (2) It defines a testable results chain: lucidity → purpose → self- actualization → social cohesion → positive social impact, with an externally validated instrument for each link. (3) It states the inversion of the valuable profile: dispersed, transversal and self-taught profiles, penalized by specialization-based selection, become the most valuable once specialized execution is performed by agents. (4) It describes a virtuous circle of bidirectional learning: people learn from multi-agent systems to work as interconnected agents without losing individuality; systems learn from community deliberation, with affected parties present and interests declared, a holistic judgment of the common good that preference data cannot capture. (5) It proposes computational linguistics of cohesion as an intersection field between cognitive science, natural language processing and applied ethics. The paper situates itself academically (§0), translates its constructs into the vocabulary of AI-and-work research funders and of corporate social responsibility and ESG reporting (§15), and states the value it offers to employers, consultancies and AI developers (§16). It deploys one method in three contexts — work, academia and science — and lists eleven falsifiable hypotheses. Limitations, including the author’s conflict of interest and instrumental circularity, are declared.