Toward AI-Native Integrated Sensing and Communication for 6G Networks

Integrated sensing and communication (ISAC) is widely regarded as a key enabling technology for sixth-generation (6G) networks. By reusing spectrum, hardware, waveforms, and computing resources, ISAC can provide environmental awareness to support emerging applications such as autonomous transportation, industrial automation, and intelligent Internet-of-Things (IoT) systems. However, conventional model-driven ISAC designs are increasingly challenged by dynamic propagation environments, heterogeneous sensing data, high-dimensional joint optimization, and stringent real-time requirements. In this context, artificial intelligence (AI), with its powerful capabilities in learning, reasoning, and prediction, is expected to play a pivotal role in the future evolution of ISAC. Nevertheless, most existing AI-enabled ISAC models are developed for isolated tasks and deployed with limited cross-module coordination, multimodal knowledge sharing, and operational feedback. To address these limitations, this article introduces AI-native ISAC as a system-level architectural paradigm for future 6G networks. The proposed framework consists of five interconnected layers, namely the physical, representation, learning, decision, and feedback layers, which jointly support multimodal environmental representation, continual model evolution, task-oriented decision-making, and closed-loop adaptation. Furthermore, we discuss the key challenges associated with the evolution from AI-assisted ISAC to AI-native ISAC, thereby providing a roadmap for future research and development.