4 ms·

AI is transforming peer review – and many scientists are worried

• MgbArapai 2y ago
  Interesting to see the mainstream recognition of how much LLMs are creeping into the peer review process. From what we’ve observed at https://papercheck.ai/en https://papercheck.ai/en where we build AI tools to assist with academic writing many researchers are already using models like GPT-4 not just for writing, but also for sanity checks on argument structure, clarity, and tone. The jump to peer review was inevitable. The ethical concerns raised are valid, especially around confidentiality and attribution. That said, AI can also act as a second pair of (very fast) eyes, catching logical gaps, passive voice, or vague claims that sometimes slip through human-only reviews. In our experience, it’s most useful when paired with human expertise not as a replacement. Offline LLMs and more transparent workflows (e.g., disclosing AI use in reviews) could help preserve trust while still evolving the process. Long-term, we may even see journals adopting AI as a formal co-reviewer tier like a spell-checker for logic and argumentation. Curious if others here are openly incorporating AI into their review process or if it’s still more of a quiet assistant?