ChatGPT produces a distinctive fingerprint, and GPTZero's classifier was trained on examples that include exactly that fingerprint. Here is the specific workflow that combines ChatGPT-aware substitution with the patterns GPTZero most aggressively flags.
Why this combination needs its own workflow
GPTZero's verdict comes with a short explanation in plain language: this passage has too-low perplexity, this paragraph has uniform burstiness. GPTZero looks at per-sentence perplexity, not just document-level. A document with a few high-perplexity sentences and many low-perplexity ones can still score human if the variance is high enough.
ChatGPT defaults to a recognizable rhythm: hook-setup-explanation paragraphs, three or five bullets per list, parallel construction across items, and a vocabulary stocked with delve, navigate, leverage, and tapestry. GPT-4o leans heavily on em dashes. Three em dashes in one paragraph is a strong AI signal in 2026.
The combination matters. A generic humanizer can move some signals, but for general writing or journalism the high-leverage moves are different. You need to strip ChatGPT's vocabulary and rhythm AND you need to specifically target the things GPTZero weights heaviest.
ChatGPT's signature words to remove
What GPTZero weights heaviest
| Pattern | Why it gets flagged | Severity |
|---|---|---|
| Low document-level perplexity | Heavy weight in GPTZero's training | Very high |
| Low burstiness (uniform sentence length) | Direct signal of model output | High |
| Repeated sentence structures | Easy to spot in general writing or journalism | High |
| Predictable next-word choices | Combines with ChatGPT's rhythm | Medium |
Concrete example
The 5-step workflow
Related guides
Run your ChatGPT draft through the humanizer
Tuned for general writing or journalism. No signup, no word limit.
Open the humanizer