As 5th Generation (5G) and Beyond 5G (B5G)networks evolve, dynamic resource allocation and managementis crucial for supporting the diversity of devices and the mixedtraffic types. Network slicing enables the logical segmentationof an infrastructure to meet specific Quality of Service (QoS)requirements posed by applications, but factors such as fluctuatingtraffic, user mobility, and cross-slice interference, posechallenges towards a proactive resource allocation. Traditionalmethods struggle with these factors, leading to inefficiencies.Therefore, this paper explores the concept of AI-driven networkperformance prediction and resource allocation framework usingTemporal Graph Networks (TGNs). By integrating TGN withthe NS-3 simulator, the work in the paper demonstrates anefficient approach to predict throughput in a network. Theproposed solution advances spatiotemporal Artificial Intelligence(AI) techniques enabling more accurate network prediction andadaptive resource optimization, towards dynamic network slicing.