Expired50% confidenceFactExact time
When a model references five samples from an author and imitates their style, the detection capability of AI text detectors drops significantly.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
7/23/2026
First Seen
This claim has expired. It is kept for the record and no longer describes the current state.
Valid until: 10/21/2026(expired)
Sources
AI日报：平头哥WAIC开源SAIL软件栈，多款大模型密集发布
Related Claims
UnverifiedUsers only need to lightly paraphrase AI output for detection scores to drop significantly.73% similarUnverified实验数据显示，经过5-10轮迭代训练后，模型在低频词汇和长尾知识上的覆盖率下降超过40%，而输出文本在表面流畅性上几乎没有可察觉的退化71% similarUnverifiedAI batch writing exhibits a quality decay phenomenon, manifesting as subsequent articles showing a noticeable gap compared to the first article in both length and depth.67% similarUnverified牛津大学2023年的一项研究量化了模型坍缩效应，仅经过五代自我训练，文本多样性指标下降了约40%65% similarUnverifiedAdding human review at critical nodes can reduce AI coding error rates by more than 60%.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/597245API
curl https://kongchang.com/api/v1/knowledge/claims/597245MCP
get_claim(id=597245)