Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Agentifying Agentic AI

Virginia Dignum, Frank Dignum

TL;DR

This paper argues that achieving genuinely agentic AI requires more than scalable learning: it needs explicit, verifiable models of cognition, interaction, and governance grounded in the Autonomous Agents and Multi-Agent Systems (AAMAS) tradition. It surveys core AAMAS concepts—BDI architectures, formal communication, mechanism design, multi-agent planning, negotiation, norms, trust, social choice, and theory of mind—and shows how they can address reliability, grounding, long-horizon reasoning, evaluation, and accountability gaps in current LLM-based agents. By proposing a synthesis of foundation-model adaptability with structured, socio-technical frameworks, the work envisions agentic systems that are not only capable and flexible but also coherent, cooperative, and socially legitimate. The practical impact is a pathway to hybrid architectures that support transparent reasoning, verifiable coordination, and governance-aligned autonomy in real-world multi-agent ecosystems.

Abstract

Agentic AI seeks to endow systems with sustained autonomy, reasoning, and interaction capabilities. To realize this vision, its assumptions about agency must be complemented by explicit models of cognition, cooperation, and governance. This paper argues that the conceptual tools developed within the Autonomous Agents and Multi-Agent Systems (AAMAS) community, such as BDI architectures, communication protocols, mechanism design, and institutional modelling, provide precisely such a foundation. By aligning adaptive, data-driven approaches with structured models of reasoning and coordination, we outline a path toward agentic systems that are not only capable and flexible, but also transparent, cooperative, and accountable. The result is a perspective on agency that bridges formal theory and practical autonomy.

Agentifying Agentic AI

TL;DR

This paper argues that achieving genuinely agentic AI requires more than scalable learning: it needs explicit, verifiable models of cognition, interaction, and governance grounded in the Autonomous Agents and Multi-Agent Systems (AAMAS) tradition. It surveys core AAMAS concepts—BDI architectures, formal communication, mechanism design, multi-agent planning, negotiation, norms, trust, social choice, and theory of mind—and shows how they can address reliability, grounding, long-horizon reasoning, evaluation, and accountability gaps in current LLM-based agents. By proposing a synthesis of foundation-model adaptability with structured, socio-technical frameworks, the work envisions agentic systems that are not only capable and flexible but also coherent, cooperative, and socially legitimate. The practical impact is a pathway to hybrid architectures that support transparent reasoning, verifiable coordination, and governance-aligned autonomy in real-world multi-agent ecosystems.

Abstract

Agentic AI seeks to endow systems with sustained autonomy, reasoning, and interaction capabilities. To realize this vision, its assumptions about agency must be complemented by explicit models of cognition, cooperation, and governance. This paper argues that the conceptual tools developed within the Autonomous Agents and Multi-Agent Systems (AAMAS) community, such as BDI architectures, communication protocols, mechanism design, and institutional modelling, provide precisely such a foundation. By aligning adaptive, data-driven approaches with structured models of reasoning and coordination, we outline a path toward agentic systems that are not only capable and flexible, but also transparent, cooperative, and accountable. The result is a perspective on agency that bridges formal theory and practical autonomy.

Paper Structure

This paper contains 10 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Explicit Architecture.
  3. Communication Protocols.
  4. Mechanism Design and Incentive Alignment.
  5. Multi-Agent Planning and Coordination.
  6. Negotiation and Argumentation.
  7. Norms and Institutions.
  8. Trust and Reputation.
  9. Collective and Strategic Decision-Making.
  10. Theory of Mind and Belief Modeling.

Figures (1)

  • Figure 1: Conceptual comparison between classical AAMAS foundations and emerging Agentic AI paradigms. The figure summarizes how established research in Autonomous Agents and Multi-Agent Systems (AAMAS) provides structured and interpretable foundations for the design of responsible and cooperative "agentic" AI systems. Each row contrasts a well-understood AAMAS concept with its current realization, or absence, in contemporary Agentic AI approaches.