Ghost Intent describes a structural phenomenon emerging in GenAI-assisted software development where executable artifacts remain behaviorally coherent while the engineering intent that produced them becomes unrecoverable. This work introduces Ghost Intent as the experiential manifestation of Authorless Traceability Collapse (ATC), a condition in which traditional software traceability assumptions fail because code generation is performed by stochastic models rather than sustained human authorship. The paper analyzes how this condition produces a "debug cost inversion", where generation effort decreases while comprehension and diagnostic effort grow non-linearly. Rather than proposing tools or workflows, the manuscript defines Ghost Intent as a governance and organizational risk within AI-native software development lifecycles. This Zenodo record archives the first stable manuscript version of the paper together with its version history artifact. Contents:- Ghost Intent manuscript (PDF)- version history document- supporting metadata
Spark Tsai (Thu,) studied this question.