English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

2.83 Million Characters Tested: Most Beliefs About the 'AI Tone' in Chinese Writing Are Wrong

Forum topic · 小凯 · 2026-08-30

Summary

An open-source Chinese project, lieflat-less-ai-tone, used a controlled corpus of 2.83 million characters—300 AI-generated articles (1.18M chars) from five major models (Claude, DeepSeek, Gemini, GPT, Kimi) versus 329 high-quality human articles (1.65M chars)—to statistically verify what actually signals AI-written Chinese text. Applying a frequency-ratio test (R = AI frequency / human frequency, threshold R ≥ 2.0), the study overturned 15 popular misconceptions and confirmed 11 genuine markers. Surprising findings: humans use metaphors 2.4× more than AI, and rhetorical questions 17× more; sentence-length uniformity shows no significant difference. The real AI tells are structural: paragraph-opening comments without referents (4.4×), reversal/contrast framing ('not X, but Y', 3.4×), lofty anthropomorphic metaphors (7.3×), empty colon-intro sentences, dash overuse (a DeepSeek/Claude habit, 47× spread across models), filler openers like 'simply put', ordinal-number headings, mirrored adjacent sentences, translation-style grammar, and replacing concrete data with vague summaries. The project converts these findings into an agent rewriting skill governed by three rules: whitelist-only edits, strict information conservation, and machine-locatable syntactic operators.

2.83 Million Characters Tested: Most Beliefs About the 'AI Tone' Are Wrong

Like watermelon buyers trusting folk wisdom, Chinese writers have long relied on unverified rules to spot AI-written text. The open-source project lieflat-less-ai-tone put these beliefs to the test with a controlled corpus of 2.83 million Chinese characters — and debunked most of them.

Experiment Design

  • AI-side corpus: 300 articles / 1.18M chars from 5 major models (Claude, DeepSeek, Gemini, GPT, Kimi), on 38 identical topics, with no retrieval or style instructions
  • Human-side corpus: 329 publicly published high-quality articles / 1.65M chars, checked for cross-source consistency
  • Verification method: frequency ratio R = generation frequency ÷ human frequency. A feature counts as an AI marker only if R ≥ 2.0 and is stable across models and human groups.
  • Result: 11 core features passed; 15 popular beliefs were refuted.

    Four Popular Myths Debunked

    1. "AI overuses metaphors" — Humans use metaphor markers 2.4× more than AI; metaphor-only paragraphs are 8× more frequent in human writing. The real difference is tenor choice: AI prefers lofty abstract figures (R = 7.3, e.g., "like a wise mentor"), humans prefer concrete life-based ones ("like a mechanic", "like a doctor friend with many WeChat contacts"). 2. "Rhetorical questions betray AI" — Humans pose questions in body text 17× more often than AI (R = 0.05). Deleting all question marks makes text *more* machine-like. 3. "AI sentence lengths are too uniform" — An artifact of early detection scripts including Markdown tables/lists. After correction, coefficients of variation were 0.58 vs 0.67 (R = 0.87) — no meaningful difference. 4. "Add function words (就/很/了) to humanize" — Artificial insertion fails verification and damages style.

    The 11 Real AI Markers

    Mostly discourse- and structure-level flaws:

    | # | Feature | R | |---|---------|---| | 1 | Paragraph-opening comments with zero referent ("Notably, …") | 4.4× | | 2 | Reversal/contrast framing ("not X, but Y") | 3.4× | | 3 | Anthropomorphic lofty metaphors | 7.3× | | 4 | Empty colon-intro sentences ("mainly as follows:") | 3.8–9.4× | | 5 | Dash overuse (DeepSeek/Claude habit) | 3.0× | | 6 | Filler openers ("simply put", "in short") | 3.2× | | 7 | Ordinal-number subheadings | 3.1× | | 8 | Structurally mirrored adjacent sentences | 2.0× | | 9 | Five typical translation-ese patterns | 2.6–5.3× | | 10 | Over-dense enumerations with pause marks | 1.8× | | 11 | Vague summaries replacing concrete data (reverse: 0.35×) | 0.35× |

    Model Styles Differ Dramatically

    There is no single "AI tone":

  • Dash frequency: DeepSeek (5.16/1000 chars) > Claude (4.25) ≫ GPT (0.11) — a 47× spread
  • Contrast framing: GPT (1.26) > DeepSeek (0.86) > Claude (0.61) > Gemini (0.29)
  • Question subheadings: Gemini (0.173) ≫ Claude (0.043) > GPT (0.008) > Kimi (0.000)
  • De-AI-flavor rules must therefore extract cross-model commonalities rather than generalize from one model.

    Three Rewriting Rules

    The project turns findings into an agent rewriting skill with three red lines:

    1. Whitelist principle: only edit the 11 verified issues; preserve all other text verbatim 2. Information conservation: never add facts/data or alter the author's judgment strength (e.g., "may improve" must not become "improves") 3. Excluding subjective semantics: rules must be machine-locatable literal and syntactic operators

    Conclusion

    De-AI-ing is not about faking human quirks — it is about removing the logical laziness, empty boilerplate, and residual grammar glitches that emerge from probabilistic prediction. The result is writing with real structure, warmth, and force.

    Project Metadata

  • Project: Lieflat Less AI Tone
  • Corpus: 629 articles, 2.827M chars, 95K sentences, 45K paragraphs
  • Models tested: Claude Opus 4.6, DeepSeek V4 Pro, Gemini 3.1 Pro, GPT-5.6 Sol, Kimi K3
  • GitHub: https://github.com/larashero3-dotcom/lieflat-less-ai-tone
  • MoxtHub: https://moxt.ai/zh-CN/hub?type=skill&id=lieflat-less-ai-tone

Tags

#ai-writing#corpus-linguistics#natural-language-processing#chinese-nlp#text-detection#writing-skills#open-source#large-language-models

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634225