Beyond Automation: Embracing Symbiotic AI for Meaningful Human Agency and Skill Preservation

Abstract

As Artificial Intelligence systems evolve into proactive, multi-agent entities, the traditional paradigm of automation experience faces unprecedented challenges regarding transparency, human agency, and long-term skill degradation. This position paper argues that moving from general Human-Centered AI (HCAI) toward Symbiotic AI (SAI) is essential for designing effective and safe agentic systems. SAI has been recently proposed as a sub-field of HCAI: it addresses systems letting humans and AI collaborate and co-evolve through mutual learning and continuous interaction. By synthesizing findings from a prior comprehensive survey and two medical case studies (on rhinocytology and on brain oncology), in this paper we discuss how SAI addresses the core themes of AutomationXP26: (1) enhancing multi-agent transparency through explainable interfaces, (2) ensuring humans remain ‘in control’ even in complex automated workflows, and (3) sustaining human skills by fostering a collaborative environment where AI acts as a partner rather than a black-box oracle. In doing so, this paper positions SAI as a design paradigm for agentic automation, offering workshop participants actionable concepts, design tensions, and empirical case studies to reason about how agentic automation can preserve human agency, transparency, and skills in real-world settings.

Full citation