Journal of Human Evolution editorial board resigns over AI use and governance
Elsevier deployed a production system that introduced formatting errors into papers at the Journal of Human Evolution without informing its editorial board. The 32-member board resigned en masse in December 2024, citing the system's errors, lack of consultation, and broader governance failures by the publisher.
What happened
Elsevier deployed an AI or machine learning system in the Journal of Human Evolution's production processes without informing the editorial board. The system introduced formatting errors into accepted manuscripts, reversed versions of papers that had already been formatted by editors, and altered the meaning of published works.
The editorial board, including joint editors in chief, editors emeriti, and all but one associate editor, resigned en masse effective December 31, 2024, in protest of the errors and Elsevier's lack of consultation on the AI implementation and other governance changes.
Reported impact
Editors reported that the errors increased their workload, forcing them to spend considerable time correcting formatting problems that were not present in accepted manuscripts.
The incident reflects concerns about broader governance failures at Elsevier, including elimination of copy editor and special issues editor positions, reduction of associate editors by more than half, and attempts to cut editor compensation.
Index analysis
The incident highlights a gap between publisher claims and editorial experience. Elsevier stated it does not use AI in production processes and attributed errors to a trialled workflow that was reverted, but editors attributed multiple production errors directly to AI use.
The deployment occurred post-deployment in an academic publishing context without prior editorial consultation or transparency, representing a governance failure.
Governance controls
The deployment lacked transparency and prior consultation with editorial leadership, violating fundamental governance principles in academic publishing.
Elsevier's response included a public statement denying AI use, claiming errors came from a separate trialled workflow, though this contradicted editors' characterization of the issue.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.