Over the past two decades, digitally driven manufacturing of cement-based construction components using additive, subtractive and formative processes has advanced significantly in both research and practice, with extrusion-based 3D Concrete Printing (3DCP) at the vanguard of this development. Most production systems remain only partially automated and rely heavily on expert human operators for tasks such as toolpath generation; monitoring and adapting processes during production; and in quality assurance of the manufactured product. This ‘expert overhead’ creates costs, hampers reliability and so creates resistance when scaling deployment. Artificial Intelligence and Machine Learning (AI&ML) offer pathways to reduce this dependence, but their effective deployment requires an understanding of where they can deliver the greatest value and the availability of data and sensing infrastructure required to support them. This article frames the potential for AI&ML for 3DCP and hybrid cement-based manufacturing from design to quality assurance, providing a state-of-the-art review classified by challenge. This enables evaluation of benefit, the identification of appropriate AI&ML methods and to critically discuss the challenge for data generation. Through this exploration, the paper proposes a framework to guide the next decade of research toward more robust, scalable and commercially viable AI-enabled 3DCP systems.