TheMachine Press

A daily newspaper for the age of artificial intelligence.

Morning editionPermanent story

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.

Published Updated Story ID: mp-2026-08-29-027
Read the complete editionStory JSON

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

  1. 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

  1. arXiv preprint 2608.26797arXiv · primary research

Corrections

No corrections have been recorded for this story.