From Automation to Authority: Introducing CCALS and REACT-HX, a Human-Expertise Framework for Safeguarding Agency in AI-Mediated Societies

This exploratory paper examines how the growing integration of artificial intelligence (AI) into workplaces, platforms, and public institutions reshapes the distribution of expertise, institutional agency, and human participation within socio-technical systems. While existing research on AI and labour markets highlights productivity gains and relatively modest net employment effects, such perspectives often overlook how AI infrastructures can redistribute discretion, accountability, and decision-making authority even when employment remains stable. Drawing on literature from labour economics, CSCW, HCI, organisational studies, and AI governance, we argue that tools introduced as assistive technologies can progressively assume greater authority in coordinating work and shaping institutional decisions. To address this shift, we introduce REACT-HX, a conceptual framework identifying five structural dimensions for sustaining meaningful human involvement in AI-mediated environments: Role Reflexivity (RR), Employment Resilience (ER), Agency Preservation (AP), Citizen-Impact Awareness (CIA), and Technology-Human Synergy (THS). We also propose Citizen-Centred AI Labour Studies (CCALS), a research agenda that reframes AI-mediated work through the lens of citizenship, emphasising that workers, IT professionals, managers, and service recipients retain the right to understand, contest, and oversee institutional decision-making. Together, REACT-HX and CCALS provide a foundation for analysing and designing AI-mediated systems that preserve human agency, professional expertise, accountability, and institutional legitimacy.