Network slicing, an emerging paradigm in mobile networks, leverages Network Function Virtualization (NFV) to enable the creation of multiple virtual networks, known as slices, over a shared physical network infrastructure. This allows operators to allocate dedicated resources and customized functions to each slice, meeting the diverse and stringent requirements of modern mobile services. Managing these functions and resources within a network slicing environment is a challenging task, often requiring real-time, efficient decision-making across all network levels. Integrating Artificial Intelligence (AI) into the network can address this challenge. This paper outlines a general framework for AI-based network slice management, introducing AI into various phases of the slice lifecycle, encompassing admission control and dynamic resource allocation within the network core and at the radio access layer. The authors suggest that a judicious application of AI for network slicing can yield significant benefits for operators, with expected performance gains ranging from 25% to 80% in representative case studies.