Always-on Personal AI for 6G networks

Integrated Artificial Intelligence (AI) and Communication has gained recognition as a central area of research. Numerous explorations regarding use scenarios and frameworks have been conducted. The current AI service NFs in 5G networks can provide terminal-oriented analyses, including UE (User Equipment) mobility analytics, UE communication analytics and movement behaviour analytics, although the data mainly comes indirectly from the network NFs and is not comprehensive or timely. The aspects related to the UE itself, such as how it acquires AI services from the networks and, in turn, how the network AI service directly utilizes data from the UE, remain largely unexplored. This paper aims to examine these aspects in depth. We propose network AI solutions, with a focus on network AI for UE and UE for network AI perspectives. We introduce "Personal AI" functionality, comprising of user connectivity services, user information services, and AI/ML (artificial intelligence and machine learning) services. We then demonstrate the solution's advantages and features through three key procedures, exhibiting the practical and effective network AI service capabilities of the solution.