This paper explores the incorporation of Machine Learning (ML) in AI art, focusing on how artists create ambiguity through the ML process. The authors analyze nine AI artworks that use computer vision and image synthesis, revealing that artists not only work closely with the ML process but also develop techniques to evoke the "ambiguity of processes." They argue that the current conceptualization of ML as a design material should be reframed to include the ML process as design elements, rather than just technical details. The paper extends HCI theories by introducing the concept of "ambiguity of process" and challenges common assumptions about ML uncertainty, dependability, and explainability. It suggests that designers should engage more closely with technical teams and consider the ML pipeline as a design space, drawing inspiration from AI artists who creatively use the ML process to create ambiguity. The findings highlight the importance of recognizing the inherent uncertainty in ML processes and how it can be leveraged to create rich interpretations and multiple meanings in AI art.This paper explores the incorporation of Machine Learning (ML) in AI art, focusing on how artists create ambiguity through the ML process. The authors analyze nine AI artworks that use computer vision and image synthesis, revealing that artists not only work closely with the ML process but also develop techniques to evoke the "ambiguity of processes." They argue that the current conceptualization of ML as a design material should be reframed to include the ML process as design elements, rather than just technical details. The paper extends HCI theories by introducing the concept of "ambiguity of process" and challenges common assumptions about ML uncertainty, dependability, and explainability. It suggests that designers should engage more closely with technical teams and consider the ML pipeline as a design space, drawing inspiration from AI artists who creatively use the ML process to create ambiguity. The findings highlight the importance of recognizing the inherent uncertainty in ML processes and how it can be leveraged to create rich interpretations and multiple meanings in AI art.