Researchers at Queensland developed a machine‑learning tool that identifies suspicious publications and mapped a broad spread of likely paper‑mill outputs affecting cancer research literature. The tool flags textual, image and metadata patterns indicative of manufactured or low‑integrity manuscripts, offering journals and funders a means to triage investigations. Authors demonstrated the utility of automated detection to surface clusters of suspect papers across publishers and disciplines, emphasizing cancer science as a revealed target. The work provides evidence for publishers and institutions to scale integrity‑checks and retraction workflows. Stakeholders in biotech—drug developers, translational labs and regulators—should consider the tool’s outputs when interpreting preclinical literature and validating reproducibility for target selection and lead optimization.