AI summaryⓘ
The authors explain that as AI moves from just understanding information to making decisions and acting on its own, traditional ways of wireless communication, which focus on sending exact bits, don't fit well anymore. They propose a new system called token communications (TokCom) that treats tokens, the basic parts used by AI language models, as the main units for exchanging information in future 6G networks. The article looks at how current communication setups need to change for TokCom, the challenges involved, and ways to solve them. They also provide an example showing how sharing tokens across different language models can work effectively. Finally, the authors suggest areas for future research to build AI-friendly communication systems based on tokens.
Artificial Intelligence (AI)Wireless CommunicationsShannon ParadigmTokensLarge Language Models (LLMs)6G NetworksToken Communications (TokCom)Communication ArchitectureHeterogeneous Language ModelsAI-native Communication
Authors
Yaru Fu, Liang Ji, Sabita Maharjan, Tony Q. S. Quek
Abstract
As artificial intelligence (AI) evolves from static perception to generative reasoning and autonomous agency, the fundamental principles of wireless communications are undergoing a paradigm shift. The classical Shannon paradigm, centered on reliable bit-level reconstruction for users, is increasingly misaligned with an emerging scenario in which the primary users of the network are interconnected AI agents. This article introduces token communications (TokCom), a novel framework that elevates tokens, i.e., the fundamental processing units of large language models (LLMs), to first-class entities for information exchange in the sixth generation wireless cellular networks (6G). We first examine the architectural transition from conventional communication systems to TokCom and identify the key challenges in implementing this transition, along with potential solution approaches. Thereafter, we present a practical case study to demonstrate the effectiveness of token sharing among heterogeneous language models. Finally, we outline promising future research directions toward realizing an AI-native, token-driven communication paradigm suitable for 6G.