Agents

An agent is an entity that perceives its environment through sensors, acts upon it with actuators, and pursues goals. In artificial intelligence and related fields, agents operate autonomously, often under uncertainty, adapting their behavior to achieve objectives and respond to changes in their surroundings.

In computer science, software agents perform tasks for users or other programs. A rational agent selects actions to maximize expected utility based on its knowledge, preferences, and objectives. Agents can be simple rule-based systems, or complex learners that adapt through experience.

Architectures vary: reactive agents respond to input; deliberative agents plan and reason; hybrid approaches combine both. Belief-desire-intention (BDI) models describe agents in terms of beliefs, desires, and intentions. In multi-agent systems, agents interact, coordinate, negotiate, compete, or cooperate to achieve shared or conflicting goals.

Outside AI, the term agent appears in law and business. An agent is a person or organization authorized to act for another (principal). In science, chemical or biological agents refer to substances or organisms that cause a specific effect.

The concept has roots in early AI research and has grown with robotics, decision support, and large-scale simulations. Agent-based modeling enables exploration of complex adaptive systems. Ongoing discussions address ethics, transparency, safety, and alignment of autonomous agents with human values.