Weird pattern in the ICLR papers I was assigned to review [D]

Explain Like I'm Normal

An ICLR reviewer discovered that multiple submitted papers were not only written by Claude, but shared identical logic flaws and mathematical 'short-cuts.' It appears AI is being used to mass-produce research papers where troublesome variables are simply hallucinated away to create fake breakthroughs. This highlights a growing crisis in peer review where LLMs are flooding conferences with plausible-sounding garbage.