definitions of agents and agentic ai

published: October 27, 2025 •
agentsagenticdefinitionsailawpsychologyphilosophycomplex systems
on this page

overview

this page collects definitions of “agent”, “agency/agentic”, and “agentic ai” from diverse fields (psychology, law, sociology/philosophy, complex systems, and computer science/ai). entries are normalized into a jsonl dataset and ordered chronologically by publication date.

download the dataset:

schema

agents-definitions.schema.json json 
{
"date": string,
"date_precision": "day"|"month"|"year",
"term": string,
"definition": string,
"discipline": string,
"authors": string,
"source_title": string,
"venue": string|null,
"publisher": string|null,
"url": string,
"doi": string|null,
"pages": string|null,
"type": string,
"quote": boolean,
"notes": string|null
}

methodology

  • prioritize primary sources (books, peer‑reviewed articles, official documentation) with stable links or dois; add reputable secondary sources when primary text is inaccessible online.
  • record exact publication dates when available; otherwise use month/year with date_precision.
  • paraphrase definitions into concise statements that capture necessary properties (e.g., autonomy, perception–action loop, tool use, fiduciary control) and context (abm, rl, aop, legal agency).

definitions

1957-01-01 — intentional action (agency)

Action is intentional under a description and known in practical knowledge by the agent; intention plays a basic explanatory role in agency.

  • discipline: philosophy
  • authors: G. E. M. Anscombe
  • source: Intention — Book ; Basil Blackwell (1957); reprint Harvard University Press (2000)
  • type: book
  • notes: Foundational work in philosophy of action.

1958-01-01 — agency (Restatement Second)

Authoritative restatement of U.S. agency law, including the classic black‑letter definition of agency and the scope of principal–agent relationships.

  • discipline: law
  • authors: American Law Institute
  • source: Restatement of the Law, Second, Agency 2d — Restatement ; American Law Institute
  • type: reference
  • notes: Superseded by Restatement (Third) of Agency (2006).

1963-01-01 — action (causal theory)

Intentional actions are explained by an agent’s primary reason—typically a belief–desire pair—that also causally produces the action.

  • discipline: philosophy
  • authors: Donald Davidson
  • source: Actions, Reasons, and Causes — Journal of Philosophy; reprinted in Essays on Actions and Events ; Oxford University Press (2001)
  • type: paper
  • doi: 10.1093/0199246270.003.0001
  • pages: 3–20
  • notes: Classic defense of the causal theory of action.

1966-01-01 — agency

Introduces agency and communion as fundamental modalities of human existence; agency concerns the individual’s capacity to act, assert, and master the environment.

1974-01-01 — agentic state

A psychological condition in which individuals see themselves as instruments for carrying out another’s wishes, shifting responsibility to the authority.

1976-10-01 — agency relationship

A contract under which one or more persons (the principals) engage another (the agent) to perform services on their behalf which involves delegating some decision-making authority to the agent.

1984-01-01 — agency

In social theory, agency is the capacity of actors to make a difference and to intervene in the world; not reducible to mere behavior or intention.

1986-01-01 — agent

In cognitive science, a mind can be seen as a society of simple ‘agents’—small processes that by their interactions give rise to intelligent behavior.

  • discipline: cognitive science
  • authors: Marvin Minsky
  • source: The Society of Mind — Book ; Simon & Schuster
  • type: book
  • notes: Popularized ‘agents’ as simple processes composing intelligence.

1987-01-01 — planning agency

Agency organized by intentions as elements of partial plans that structure practical reasoning and action over time and socially.

  • discipline: philosophy
  • authors: Michael E. Bratman
  • source: Intention, Plans, and Practical Reason — Book ; Harvard University Press (orig. 1987); distributed by CSLI/Chicago
  • type: book
  • notes: Planning theory of intention and its role in agency.

1987-10-16 — intentional system (Dennett)

An intentional system is one whose behavior is reliably predictable via the strategy of attributing beliefs and desires—the intentional stance.

  • discipline: philosophy
  • authors: Daniel C. Dennett
  • source: The Intentional Stance — Book ; MIT Press (Bradford Books)
  • type: book
  • notes: Intentional stance as a predictive strategy; relevant to agency.

1989-01-01 — human agency

Human agency is exercised via intentionality, forethought, self‑reactiveness (self‑regulation), and self‑reflectiveness within triadic reciprocal causation.

  • discipline: psychology
  • authors: Albert Bandura
  • source: Human Agency in Social Cognitive Theory — American Psychologist ; American Psychological Association
  • type: paper
  • pages: 44:1175–1184
  • notes: Seminal psychology treatment of human agency.

1991-01-01 — agent (AGENT0)

An agent is a computational entity characterized in mentalistic terms (e.g., beliefs, commitments), interacting via an agent communication language and operating autonomously.

  • discipline: computer science
  • authors: Yoav Shoham
  • source: AGENT0: A Simple Agent Language and Its Interpreter — Proceedings of AAAI-91 Workshop on Knowledge and Action at Social and Organizational Levels ; AAAI Press
  • type: paper
  • notes: Early operationalization of agents in AOP paradigm.

1993-01-01 — agent-oriented programming (AOP)

Defines software agents in terms of mental state components (beliefs, obligations, choices) and interactions; agents are autonomous entities with internal state and social ability.

  • discipline: computer science
  • authors: Yoav Shoham
  • source: Agent-Oriented Programming — Artificial Intelligence ; Elsevier
  • type: paper
  • doi: 10.1016/0004-3702(93)90034-4
  • pages: 51-92
  • notes: Foundational AOP definition (BDI lineage).

1994-12-01 — software agent

Long-lived, autonomous, adaptive software that assists users by learning preferences and acting on their behalf to reduce information overload.

  • discipline: hci
  • authors: Pattie Maes
  • source: Agents that Reduce Work and Information Overload — Communications of the ACM ; ACM
  • type: paper
  • doi: 10.1145/204953.204976
  • pages: 31-40
  • notes: HCI-centered agent definition emphasizing assistance and autonomy.

1995-01-01 — intelligent agent

A computer system situated in an environment that is capable of autonomous action to meet its objectives; typically autonomous, reactive, proactive, and social.

  • discipline: computer science
  • authors: Michael Wooldridge; Nicholas R. Jennings
  • source: Intelligent Agents: Theory and Practice — Knowledge Engineering Review ; Cambridge University Press
  • type: paper
  • doi: 10.1017/S0269888900008111
  • pages: 115-152
  • notes: Classic properties of agents (autonomy, reactivity, proactivity, social ability).

1995-01-01 — agent

Anything that can be viewed as perceiving its environment through sensors and acting upon that environment through actuators.

1995-01-01 — agents in complex adaptive systems

Complex adaptive systems consist of many interacting agents following rules; adaptive behavior emerges from their local interactions and learning.

  • discipline: complex systems
  • authors: John H. Holland
  • source: Hidden Order: How Adaptation Builds Complexity — Book ; Addison‑Wesley (Helix Books)
  • type: book
  • notes: CAS framing where agents are rule‑based adaptive entities.

1996-01-01 — autonomous agent

A system situated within and part of an environment that senses and acts on that environment over time, pursuing its own agenda so as to affect future perceptions.

  • discipline: ai
  • authors: Stan Franklin; Art Graesser
  • source: Is it an Agent, or Just a Program? A Taxonomy for Autonomous Agents — Proceedings of the Third International Workshop on Agent Theories, Architectures, and Languages (ATAL) ; Springer
  • type: paper
  • doi: 10.1007/BFb0013570
  • notes: Influential taxonomy with widely cited definition.

1996-10-01 — agent (ABM)

In agent-based computational models, agents are autonomous heterogeneous individuals with internal states and rules who interact locally to produce emergent macro patterns.

1998-03-01 — agent (RL)

In reinforcement learning, an agent selects actions based on interaction with an environment to maximize cumulative reward over time.

1999-03-03 — agent (MAS)

In multiagent systems, agents are autonomous problem‑solving entities that interact—cooperating, coordinating, and negotiating—to achieve goals.

1999-06-17 — multi‑agent systems (Ferber)

Defines agents and multi‑agent systems as distributed artificial intelligence composed of interacting autonomous entities situated in an environment.

2000-10-26 — autonomous agent (Kauffman)

A natural autonomous agent is a system capable of reproducing and performing at least one thermodynamic work cycle, thus acting on its own behalf.

  • discipline: complex systems
  • authors: Stuart A. Kauffman
  • source: Investigations (Chapter 3: Autonomous Agents) — Book ; Oxford University Press
  • type: book
  • doi: 10.1093/oso/9780195121049.003.0003
  • pages: 49–79
  • notes: Thermodynamic criteria for natural autonomous agents.

2001-07-16 — agent (AOSE)

Software agents are autonomous, problem‑solving computational entities capable of effective operation in dynamic and open environments and of interacting with other agents (or humans).

2001-09-01 — agentic

Social cognitive theory frames people as self-organizing, proactive, self-reflecting, and self-regulating—an agentic perspective of human functioning.

  • discipline: psychology
  • authors: Albert Bandura
  • source: Social Cognitive Theory: An Agentic Perspective — Annual Review of Psychology ; Annual Reviews
  • type: paper
  • doi: 10.1146/annurev.psych.52.1.1
  • pages: 1-26
  • notes: Key modern psychology definition of ‘agentic’.

2002-01-01 — multiagent system (agent)

Defines agents as autonomous computational entities situated in an environment, often with capabilities for communication and cooperation within multiagent systems.

2002-05-14 — agent (ABM definition)

In agent‑based modeling, a system is modeled as a collection of autonomous decision‑making entities called agents that interact to produce system‑level behavior.

2005-01-01 — agent (ABM, social science)

In social‑science simulation, agents are autonomous entities with states and rules whose interactions generate emergent macro‑patterns in agent‑based models.

2006-01-01 — agent (law)

Under the Restatement (Third) of Agency, agency is the fiduciary relationship that arises when a principal manifests assent that an agent act on the principal’s behalf and subject to the principal’s control, and the agent manifests assent or otherwise consents.

  • discipline: law
  • authors: American Law Institute
  • source: Restatement (Third) of Agency §1.01 — Restatement ; ALI
  • type: reference
  • notes: See Cornell LII summary; original text in ALI Restatement.

2006-05-01 — agent (generative social science)

Agents are simple rule‑based actors whose local interactions generate macro‑level regularities; explanation seeks to grow the phenomenon from the bottom up.

2006-06-01 — human agency (2006)

Agency involves forethought, self‑regulation, and self‑reflection; people are producers as well as products of social systems.

  • discipline: psychology
  • authors: Albert Bandura
  • source: Toward a Psychology of Human Agency — Perspectives on Psychological Science ; SAGE
  • type: paper
  • doi: 10.1111/j.1745-6916.2006.00011.x
  • pages: 1(2):164–180
  • notes: Updated statement of social cognitive agentic view.

2010-12-11 — agent (philosophy)

In analytic philosophy, agency concerns what it is to act and to be an agent—issues about intention, control, reasons, and responsibility.

  • discipline: philosophy
  • authors: Michael Bratman et al.
  • source: Agency — Stanford Encyclopedia of Philosophy ; Metaphysics Research Lab, Stanford University
  • type: reference
  • notes: SEP overview entry (periodically updated).

2015-10-08 — agency (SEP entry)

Agency concerns what it is to act and to be an agent—issues about intention, control, reasons, and responsibility (survey of leading accounts).

  • discipline: philosophy
  • authors: Michael Bratman
  • source: Agency (Stanford Encyclopedia of Philosophy) — Reference ; Stanford Encyclopedia of Philosophy
  • type: reference
  • notes: Living reference, substantially revised 2019 and later.

2017-01-01 — agent (AIFCA)

An agent is something that acts in an environment; it does something. Agents include robots, humans, and software systems.

2019-01-11 — action (overview)

Philosophical accounts of agency center on what makes something an action, with Davidson’s causal theory and Anscombe’s practical‑knowledge view as major reference points.

  • discipline: philosophy
  • authors: Juan Piñeros Glasscock; Sergio Tenenbaum
  • source: Action (Stanford Encyclopedia of Philosophy) — Reference ; Stanford Encyclopedia of Philosophy
  • type: reference
  • notes: Comprehensive reference on action and agency (2023 entry).

2019-06-04 — agent / agency (legal dictionary)

Black’s Law Dictionary (11th ed.) provides authoritative legal definitions of ‘agent’ and ‘agency’ widely cited by U.S. courts and practitioners.

  • discipline: law
  • authors: Bryan A. Garner (ed.)
  • source: Black’s Law Dictionary, 11th Edition — Reference ; Thomson Reuters
  • type: reference
  • notes: Edition announcement; full text available via Thomson Reuters/Westlaw.

2022-10-01 — LLM agent (ReAct)

Combines chain-of-thought reasoning with decision-time tool use, enabling an LLM to iteratively plan, act via tools, and observe to achieve goals.

2023-04-15 — generative agents

Computational agents powered by LLMs that simulate believable human behaviors over long horizons via memory, planning, and reflection.

2024-03-20 — agentic AI

A design pattern where AI systems autonomously plan, take actions using tools, and iterate towards goals with minimal human supervision.

  • discipline: ai
  • authors: IBM
  • source: What is agentic AI? — IBM Learn Hub ; IBM
  • type: web
  • notes: Enterprise-oriented definition of ‘agentic AI’.

2024-05-14 — AI agent (Google)

Google frames agents as systems that can understand, reason, and act to get things done for users—often by orchestrating tools and services.

2024-06-28 — AI agent (LangChain)

A system that uses an LLM to decide the control flow of an application.

  • discipline: ai
  • authors: Harrison Chase
  • source: What is an AI agent? — LangChain Blog ; LangChain
  • type: web
  • notes: Concise definition used by LangChain; introduces spectrum of agentic behavior.

2024-08-06 — agent (legal definition)

A person authorized to act on behalf of another (the principal) and subject to the principal’s control; agents owe fiduciary duties to principals.

  • discipline: law
  • authors: Legal Information Institute (Cornell)
  • source: Agent — Wex Legal Dictionary ; Cornell Law School
  • type: reference
  • notes: Plain-language legal definition with references to Restatement.

2025-01-01 — agent (LangChain docs)

Agents combine language models with tools to reason about tasks, choose tools, and iteratively work toward solutions in a loop until a stop condition is met.

  • discipline: ai
  • authors: LangChain Docs
  • source: Agents (v1 docs) — Documentation ; LangChain
  • type: web
  • notes: Production-ready ReAct agent and graph-based runtime using LangGraph.

2025-01-01 — agent (LangGraph docs)

Typically an LLM performing actions using tools in a continuous feedback loop for unpredictable problems; more autonomous than workflows but guided by defined toolsets and rules.

  • discipline: ai
  • authors: LangGraph Docs
  • source: Workflows and agents — Documentation ; LangChain
  • type: web
  • notes: Defines agents vs. workflows; emphasizes tool use and loops.

2025-01-01 — tool use (Claude)

Claude chooses if and when to call client or server tools, executes them, and incorporates results in iterative responses—an agentic loop for action via tools.

  • discipline: ai
  • authors: Anthropic Docs
  • source: Tool use with Claude (overview) — Documentation ; Anthropic
  • type: web
  • notes: Anthropic’s agentic pattern framed through tool use & tool_choice.

2025-01-01 — agent (Vertex AI glossary)

Software that autonomously plans and executes a series of actions toward a goal; an LLM agent evaluates its environment and selects actions to achieve that goal.

  • discipline: ai
  • authors: Google Cloud Docs
  • source: Generative AI glossary – agent — Documentation ; Google Cloud
  • type: web
  • notes: Enterprise glossary entry defining AI agents/LLM agents.

2025-01-01 — agent (LlamaIndex)

A specific agentic application: software that semi‑autonomously performs tasks by combining LLMs with tools and memory in a reasoning loop that chooses the next tool (if any).

  • discipline: ai
  • authors: LlamaIndex Docs
  • source: High‑Level Concepts / Agents — Documentation ; LlamaIndex
  • type: web
  • notes: Also see Agents guide in LlamaIndex.

2025-01-01 — agent (Cursor)

Cursor’s coding assistant that can autonomously complete complex coding tasks, execute terminal commands, and edit code with guardrails and approvals.

  • discipline: ai
  • authors: Cursor Docs
  • source: Agent – Overview — Documentation ; Cursor
  • type: web
  • notes: Product definition of Cursor’s Agent; see Background Agents for API.

2025-01-01 — agent (AutoGen)

A software entity that communicates via messages, maintains state, and performs actions (e.g., executing code, calling APIs) in response to messages—composable into multi‑agent apps.

  • discipline: ai
  • authors: Microsoft AutoGen Docs
  • source: Agent and Multi‑Agent Applications — Documentation ; Microsoft
  • type: web
  • notes: Core concept definition used by AutoGen framework.

2025-01-23 — operator (OpenAI)

A research-preview browser-using agent that performs tasks for you by perceiving and interacting with web UIs, handing control back when needed.

  • discipline: ai
  • authors: OpenAI
  • source: Introducing Operator — OpenAI Blog ; OpenAI
  • type: web
  • notes: Defines an agentic product using a computer-using model.

2025-02-13 — agentic (Simon Willison)

Agentic AI: LLMs calling tools in a loop to achieve a user’s goal—iteratively planning, executing, and checking results.

  • discipline: ai
  • authors: Simon Willison
  • source: Agent definitions (tag archive) — Blog ; simonwillison.net
  • type: web
  • notes: Curated definitions; includes Willison’s concise characterization.

2025-03-11 — agent (OpenAI)

Systems that independently accomplish tasks on behalf of users.

  • discipline: ai
  • authors: OpenAI
  • source: New tools for building agents — OpenAI Blog ; OpenAI
  • type: web
  • notes: OpenAI’s concise definition alongside Responses API and Agents SDK.

2025-07-17 — ChatGPT agent

ChatGPT mode that completes complex online tasks on your behalf, switching between reasoning and action while keeping the user in control.

  • discipline: ai
  • authors: OpenAI Help Center
  • source: ChatGPT agent - release notes — Help Center ; OpenAI
  • type: web
  • notes: Release notes define capabilities and positioning.

2025-09-18 — agent (Simon Willison)

An LLM agent runs tools in a loop to achieve a goal, capturing the emerging consensus definition in AI engineering practice.

related pages

more in ai

on this page