This paper attempts to provide a unifying conceptual framework to interpret the perils and promises of artificial intelligence (AI) applications in criminal justice systems. As platform technologies that support a myriad of potential applications, the impact of AI systems upon criminal justice will be amplificatory, divergent, and simultaneous. A unifying framework will facilitate a holistic appreciation of why AI systems might destabilise criminal justice systems and suggest appropriate responses depending on the type of criminal legal disruption at its root. We elaborate this theoretical framework by analysing problems posed by generative “deep-fake” technology, in the context of the criminal justice system in England and Wales. The value of this framework is that it forces us to ask questions about pre-existing normative and procedural responses in a way that reveals future problems, instead of bounded discussions of the application to and adaptability of existing systems which does not.