AI-Native Networks White Paper

The convergence of AI and communications is transforming the very role and value of communication networks. While partial deployment of AI within communication networks is already underway, the transition to "AI-nativenetworks" — where the lifecycles of AI and communication networks are fused, and humans and AI collaboratively determine what is decided, where, and how, dynamically configuring decision structures according to context — represents not merely a technological advancement, but a fundamental shift in the assumptions underlying social infrastructure. Yet a framework for holistically explaining and verifying system-wide behavior remains underdeveloped.

Under this transition, three longstanding assumptions begin to erode. First, the interdependence among communication, computing, cloud, and power infrastructures makes per-facility investment evaluation inadequate(investment justification). Second, when AI-driven control spans multiple infrastructure layers, the very premise of what constitutes a single evaluable unit is called into question (evaluation and verification). Third, when human-AI role division may shift during operation, establishing regulatory accountability in advance becomes problematic(liability allocation). These are not hypothetical future concerns — they are emerging in current deployment and operational practice.

This whitepaper makes this structural transformation visible and provides a shared foundation for stakeholders with differing perspectives to engage in dialogue under common premises.