Documented Presence and Process Location of Generative Artificial Intelligence in Industrial Scam Compounds and the Limits of the Transformation Claim

Abstract Peer-reviewed research on Southeast Asian industrial scam compounds documents hierarchical, coercive organisation and the presence of generative models, synthetic media and related tools in the operational repertoire. It does not, however, isolate those tools as the measured cause of labour change, success rates or organisational form. This article, therefore, asks a more limited set of questions. First, what does the published record establish about the organisation of compound-based cyber-fraud? Second, where in the crime-commission process are generative and synthetic-media tools located, and at what level of documentation? Third, what remains unshown about transformation? The study is a structured qualitative synthesis of recent peer-reviewed work on compounds, trafficking for forced criminality, cybercrime organisation and generative fraud, supplemented by a bounded strand of official assessments and investigative process sources coded only for the described process, role or tool. Official estimates are reported as official estimates. A composite script table locates documented insertion points without treating presence as effect. Three claims are kept separate: compound organisation is strongly evidenced; generative and synthetic tools are moderately evidenced as repertoire items; AI as a cause of labour, success or form change is not shown. Agentic, multi-step orchestration is distinguished from content generation and is treated as thinly evidenced in the peer-reviewed set. The distinctive contribution is an open adjudication problem: the compound-level measures that would decide between an incremental-tool reading and a reorganisation reading.