safety
More Human Examples Made the Machine Distribution Harder to See
A theoretical analysis shows repeated paraphrasing can move machine-written responses toward an empirical human-writing distribution.
Summary
A theoretical analysis shows repeated paraphrasing can move machine-written responses toward an empirical human-writing distribution.
Under stated mixing and stability assumptions, the paper derives a convergence rate and describes how the needed human samples and paraphrasing rounds scale with the target error. The result characterizes a strategic evasion process in a controlled multi-sample setting; it does not prove that all AI text is presently undetectable.
Why it matters
A theoretical analysis shows repeated paraphrasing can move machine-written responses toward an empirical human-writing distribution.
Limits and context
- The result characterizes a strategic evasion process in a controlled multi-sample setting; it does not prove that all AI text is presently undetectable.
Key claims
A theoretical analysis shows repeated paraphrasing can move machine-written responses toward an empirical human-writing distribution.
Qualification: The result characterizes a strategic evasion process in a controlled multi-sample setting; it does not prove that all AI text is presently undetectable.
Evidence: source-2026-08-29-016
Sources
- arXiv preprint 2608.26797arXiv · primary research
Corrections
No corrections have been recorded for this story.