The AI peer review debate: is just running a paper through Claude dishonest?
anshulkundaje · x · 2026-10-29
jmschreiber91 draws a key distinction between fully AI and AI-assisted review: if you're asked to review a paper and just run it through Claude, isn't that dishonest — and couldn't journals provide auto-reviews themselves? Kundaje responds that everything should be disclosed either way, including the common-but-hidden practice of grad students/postdocs reviewing for PIs, and pure AI reviews are fine if explicitly marked.
More from AGI Musings
- NatPurser: short AI timelines may be spurring less rigorous policy ideas — NathanpmYoung · 2026-10-29
- Comparing AI hype to nanomachines vaporware 'will age poorly', argues dev — jd_pressman · 2026-10-29
- Engineer pushes back on AI job-loss narrative: big tech layoffs follow earnings cycles — Nasereliver · 2026-10-29
- Berkeley researcher on why neolabs matter: choosing the right question is innovation — BeidiChen · 2026-10-29
- Sundar Pichai invokes RSI on earnings call, expects 12 months of acceleration — kimmonismus · 2026-10-29
- Developer dumps all Apple stock, betting an AI OS replaces iOS within two years — jonathan_wilke · 2026-10-29