AI flags more than 250,000 suspicious cancer research papers
Summary
A new AI system scanned 2.6 million cancer papers from 1999 to 2024 and flagged over 250,000 with writing patterns resembling suspected paper-mill output. The result suggests industrial-scale contamination of the cancer literature rather than isolated misconduct.
Why it matters
If a large fraction of the cancer record is unreliable, downstream science and patient-facing decisions can be systematically misled.