Archive — US — techcrunch
Academics accuse AI startups of co-opting peer review for publicity
A controversy emerges at ICLR over AI-generated studies, with labs Sakana, Intology, and Autoscience drawing criticism for co-opting peer review processes. Sakana informed ICLR prior to submission, but the others did not. Critics argue this exploits unpaid labor and disrespects the effort involved in peer review, highlighting a growing concern about synthetic text in academia and the need for regulation in AI study evaluations.
Key points
- Three AI labs (Sakana, Intology, Autoscience) submitted AI-generated studies to ICLR workshops.
- Sakana obtained consent before submission; Intology and Autoscience did not.
- Critics argue this undermines peer review integrity and exploits academic labor.
- Peer review is time-consuming; academics report spending 2-4 hours per study.
- There's a rising percentage of AI-generated content in academic papers, with some estimates between 6.5% and 16.9%.
- Sakana withdrew its submission for transparency; other labs faced backlash for their methods.
Why it matters
- The increasing prevalence of AI in academia mirrors broader societal debates on automation and labor exploitation.
- 2025 AI Trends highlight the necessity for ethical guidelines in AI-generated content and evaluation.
- This incident underscores potential pressures on peer review systems amidst growing AI-generated submissions, as seen in events like NeurIPS.
The incident reflects ongoing tensions between academia and AI, emphasizing the need for ethical considerations and better regulation to preserve academic integrity in the face of advancing artificial intelligence.