Mistake: turning style into proof
Polished wording, repeated transitions and balanced paragraphs can appear in writing from many sources. Short samples provide especially little context for interpretation.
The detector describes observable language patterns. Its result is uncertain and does not establish whether a human or an AI wrote the text.
- Read the whole passage.
- Identify visible patterns.
- Avoid an authorship conclusion.
Mistake: expecting a score
This service does not calculate an AI percentage or an accuracy score. A confident-sounding explanation is still model-generated commentary that needs review.
Do not convert descriptive wording into your own numerical probability. That would create a measurement the service has not produced.
- Check what the result actually says.
- Remove invented percentages.
- Keep the uncertainty visible.
Mistake: ignoring context
Translated texts, formal notices and heavily edited drafts may use constrained wording. Compare a comment with the document purpose before deciding whether an edit helps.
For important reviews, examine sources, revision history supplied by the writer and relevant work records. Language analysis can support editing, but should not decide misconduct by itself.
- Ask about the document purpose.
- Check available source material.
- Use a fair review process.
Worked example
All paragraphs begin with a similar transition.
This is a repeated structural pattern worth editing if it distracts readers. It does not prove AI authorship.
The example shows how to keep an observable signal separate from an unsupported conclusion.
Before you finish
- No invented AI score.
- No decision from a short excerpt.
- Translation context considered.
- Evidence reviewed separately.
Questions and answers
Can I use the result as sole evidence against a writer?
No. This tool cannot reliably determine authorship.
Can I still improve the passage?
Yes. Address specific reader problems, such as repetition or vague phrasing, while preserving the facts.