Fraud, AI slop and huge profits: is science publishing broken?

Overview: The publishing backbone and the crisis

Publishing in science operates as a cornerstone of the scientific method. Papers are submitted to journals where volunteer editors and peer reviewers scrutinize the work, shaping what becomes part of the scientific record. Yet across the last decade the system has faced mounting pressure. The transcript notes that annual outputs rose from about 2 million articles in 2016 to roughly 3 million by 2022, while the number of new scientists did not keep pace, creating a time bottleneck for reading, writing, and reviewing. The result is information noise, with a flood of papers some of which are mediocre or worse, and a substantial hidden cost in time and energy. The conversation highlights the billion-dollar value of peer review in 2020 and the broader strain on researchers who volunteer as editors while juggling their own work. "the scientific publishing system and the academic job market are a bit broken" - Mark Hanson

As the volume grows, so does the pressure to publish, intensifying the “publish or perish” culture. The episode also explains special issues—collections intended to direct a field. Publishers, the speakers argue, sometimes treat these invitations as mass mailings to solicit articles, contributing to unsustainable volumes and diluting meaningful conversations. The discussion emphasizes the difference between genuine scholarly collaboration and attempts to manufacture “special” status, warning that not every field has enough truly novel discussions to justify thousands of such issues.