Socio-technical aspects of Agentic AI

TL;DR

The paper addresses the gap between technically focused agentic AI research and its societal implications by proposing a socio-technical analysis using the MAD–BAD–SAD framework. It defines agentic AI, surveys architectural and application domains, and then maps technical modules to ethical and societal concerns, emphasizing governance, accountability, and data stewardship. The main contributions include a structured MAD–BAD–SAD analysis, open challenges, and future directions that integrate socio-economic adaptation and long-term societal integration. The work highlights the need for governance frameworks, transparency, and value alignment to ensure sustainable and responsible deployment of autonomous, goal-directed AI across healthcare, education, smart cities, networking, EO, and Industry 5.0. This integrated perspective aims to inform designers, policymakers, and practitioners about balancing autonomy with safety, trust, and social welfare in real-world settings.

Abstract

Agentic Artificial Intelligence (AI) represents a fundamental shift in the design of intelligent systems, characterized by interconnected components that collectively enable autonomous perception, reasoning, planning, action, and learning. Recent research on agentic AI has largely focused on technical foundations, including system architectures, reasoning and planning mechanisms, coordination strategies, and application-level performance across domains. However, the societal, ethical, economic, environmental, and governance implications of agentic AI remain weakly integrated into these technical treatments. This paper addresses this gap by presenting a socio-technical analysis of agentic AI that explicitly connects core technical components with societal context. We examine how architectural choices in perception, cognition, planning, execution, and memory introduce dependencies related to data governance, accountability, transparency, safety, and sustainability. To structure this analysis, we adopt the MAD-BAD-SAD construct as an analytical lens, capturing motivations, applications, and moral dilemmas (MAD); biases, accountability, and dangers (BAD); and societal impact, adoption, and design considerations (SAD). Using this lens, we analyze ethical considerations, implications, and challenges arising from contemporary agentic AI systems and assess their manifestation across emerging applications, including healthcare, education, industry, smart and sustainable cities, social services, communications and networking, and earth observation and satellite communications. The paper further identifies open challenges and suggests future research directions, framing agentic AI as an integrated socio-technical system whose behavior and impact are co-produced by algorithms, data, organizational practices, regulatory frameworks, and social norms.

Figures (6)

• Figure 1: Conceptual focus of this paper, highlighting a socio-technical framing of agentic AI systems.
• Figure 2: An overview of our socio-technical taxonomy of agentic AI and a conceptual roadmap for the organization and argumentative flow of the remainder of the paper.
• Figure 3: Overview of an agentic AI architecture in which an AI agent integrates perception, cognition, and planning, memory, tools, and execution within a closed feedback loop, while higher-level mechanisms for task decomposition, orchestration, shared context, and inter-agent collaboration enable coordinated, goal-directed behavior across multiple agents.
• Figure 4: The Socio-Technical Stack of agentic AI. \ref{['fig4a']} Recent evolution of agentic core and its infrastructure and technical context. \ref{['fig4b']} The layered architecture of agentic AI, integrating technical and social dimensions. The central core represents the technical modules while the surrounding outer layer denotes the socio-technical constraints and enablers that shape and regulate each module.
• Figure 5: The MAD–BAD–SAD Intertwined flow. Illustration of how technical capabilities in agentic AI, such as autonomous planning and execution, interact with social environments to produce broader societal outcomes, and how technical design choices propagate into moral, safety, and adoption-related challenges.