The world of academic publishing is undergoing a significant transformation with the integration of artificial intelligence (AI). This story, however, is not about the benefits of AI but rather the potential pitfalls and the human cost of its implementation.
The Resignation of a Neuroscience Editor
Michael Okun, an associate professor of neuroscience, has stepped down from his role as an associate editor at Frontiers in Systems Neuroscience. His resignation is a protest against the journal's AI system, which he believes is undermining the very foundation of academic integrity.
Okun's experience highlights a critical issue: the journal's AI system, Artificial Intelligence Research Assistant (AIRA), appears to be making decisions that should be left to human editors. In his case, AIRA contacted potential reviewers without Okun's consent, and then proceeded to revoke his own invitations to experts in the field.
AIRA's Impact on Editorial Control
What makes this particularly fascinating is the power dynamic at play. Okun argues that the journal's focus on quantity over quality is evident in AIRA's design. The system seems to be intentionally removing human editors from the equation, raising questions about the journal's priorities.
In my opinion, this is a crucial moment for the academic community to reflect on the role of AI. While automation can streamline processes, it must not compromise the core values of academic publishing: integrity, expertise, and human judgment.
The Journal's Response
Frontiers' chief executive editor, Frederick Fenter, acknowledges Okun's bad experience and invites further discussion. He emphasizes the ability of editors to pause reviewer invitations, a feature that was not clear to Okun.
Fenter's statement also highlights the journal's commitment to saving scientists' time, which is a noble goal. However, it seems that in this case, the focus on efficiency has led to a breakdown in communication and potentially compromised the quality of the peer review process.
Other Editors' Experiences
While some editors, like Shuzo Sakata, have had similar negative experiences with AIRA, others have not encountered the same issues. This suggests that the problem may lie in the system's design or implementation, rather than a universal issue with AI-assisted editing.
The Broader Implications
The use of AI in academic publishing is a double-edged sword. On one hand, it can expedite processes and reduce the burden on human editors. On the other, it can lead to unintended consequences, such as the revocation of expert invitations or the selection of reviewers with limited expertise.
What many people don't realize is that AI, despite its sophistication, lacks the nuanced understanding and judgment that human editors bring to the table. It's this human element that ensures the quality and integrity of academic research.
A Way Forward
Okun offers a balanced perspective, suggesting that the use of large language models to find experts can be beneficial as long as human editors are consulted. Alternatively, he proposes a fully automated system, provided publishers are transparent about AI's role.
This raises a deeper question: can we strike a balance between the efficiency of AI and the expertise of human editors? It's a challenge that the academic community must address as AI continues to shape the future of scholarly publishing.
