DocAI: A Framework for Generating and Identifying AI-Edited Documents

Document-level detection of AI-generated content is largely unaddressed, and off-the-shelf tools do not fill the gap: across nine AI-text and AI-image detectors, and both ways of composing them—score the whole page, or segment it and score each region—no configuration exceeds AUROC $0.66$. The task demands dedicated tools and the data to build them; we provide both. DocAIgen edits real documents into large-scale corpora with region-level ground truth, from which we release a $42,503$-page benchmark. DocAIdet is a document-native Vision-Language Model that reads the page as a whole, jointly detecting and segmenting AI-generated regions. It reaches page-level AUROC $0.944$ in-generation and $0.929$–$0.970$ on two unseen generators ($0.999$ at document level), with Dice $0.895$ and $0.699$–$0.863$ respectively; the strongest piecewise baseline, retrained on our own data, reaches $0.874$ AUROC and $0.270$ Dice on the harder unseen generator, and the margins are significant with non-overlapping $95\%$ intervals. We release the engine, the benchmark and the detector.