swarm

a.k.a. AI swarm, agent swarms, swarm chaser

An AI swarm is where multiple autonomous or semi-autonomous artificial intelligence agents collaborate to accomplish a shared goal. Instead of relying on a single model to handle an entire request from start to finish, a swarm breaks down complex workflows and assigns specialized roles to multiple AI agents. 

How an AI swarm works: 
Roles: Each AI agent in the swarm has a specific specialty, instruction set, and set of tools (such as research, coding, writing, or database querying) so they distribute the work.
Coordination: AI agents communicate with each other, handoff tasks back and forth, share context or memory, and adapt to intermediate results in real time.
Styles: AI swarms can operate via a master agent (centralized), a hierarchical structure (architect), or peer-to-peer interactions (decentralized) inspired by biological systems like insect colonies. 

Historical perspective: As of 2026 swarms were real and AI engineers said that next-generation AI agent swarms could seize control of the internet, banking systems, or the electrical grid, or trick nations into responding to a fake nuclear attack. Bad actors could use them to create biological weapons. Bill Gates said that if AI’s development is not slowed and thoughtfully regulated, it will certainly be powerful enough to cause a billion deaths. Film at 11. 

NetLingo Classification: Online Jargon

Updates