Beyond A2A and MCP: How LOKA’s Universal Agent Identity Layer changes the game
1 min read
Summary
Researchers from Carnegie Mellon University have proposed Layered Orchestration for Knowledgeful Agents (LOKA), a new interoperability protocol designed to govern the identity, accountability and ethics of AI agents.
As AI becomes more ubiquitous, the team stated in a paper that there is an increasing need for a standardised framework to govern interactions, with existing systems lacking a uniform protocol for communication, ethical reasoning and compliance with jurisdictional regulations.
LOKA aims to address these gaps through a “foundational architecture” that enables agents to prove their identity, exchange “ethically annotated messages” and establish ethical governance, with each agent assigned a unique, cryptographically verifiable identifier.
The protocol is designed as a stack of four layers: identity, communication, ethics and security.
The ethics layer uses collective decision-making models to enable agents to determine their next steps whilst adhering to ethical standards, with the security layer using “quantum-resilient cryptography” to ensure secure communications.