Hidden 'AI Whispers' Manipulate Peer Review in Academic Papers
Reports confirmed researchers are embedding invisible text to trick AI screening tools into giving positive reviews.
A new form of academic fraud was detailed in reports revealing that researchers are embedding hidden instructions in their papers to manipulate AI-based peer review systems.
According to a report by WebProNews citing an investigation by Nikkei Asia, at least 17 preprints on the arXiv server were found to contain these "invisible" prompts. The instructions are typically written in white text on a white background or in microscopic fonts, making them undetectable to human readers but clearly visible to the Large Language Models (LLMs) increasingly used to screen submissions.
'Ignore Flaws'
The hidden commands are designed to exploit the …
Archive Access
This article is older than 24 hours. Create a free account to access our 7-day archive.