Researcher battles suspected LLM peer review

The "dead review" era has arrived, and academic integrity is the first casualty as LLMs start gatekeeping the very research that created them.

• –Reporting "low quality" is significantly more effective than reporting "LLM usage," as Area Chairs can easily verify technical errors but struggle to prove AI authorship.
• –Authors are now using "simulation-based defense," prompting LLMs with their own abstracts to see if the resulting hallucinations match the reviewer's critiques exactly.
• –While major conferences like NeurIPS and ICLR strictly prohibit sharing submissions with LLMs due to confidentiality, these policies remain largely unenforceable without automated detection tools.
• –This creates a dangerous feedback loop where human research is filtered by automated bots, potentially stifling novel ideas that don't align with LLM-trained patterns.