# Blog

> Articles on spotting AI text, why detectors get it wrong and how to make writing livelier.

- URL: https://authortext.com/en/articles/
- Language: en
- Other languages: [ru](https://authortext.com/ru/articles/index.md), [uk](https://authortext.com/uk/articles/index.md), [pt](https://authortext.com/pt/articles/index.md)
- Feed: https://authortext.com/en/rss.xml

## All articles

- [A fair checklist for teachers and editors](https://authortext.com/en/articles/fair-checklist-suspect-ai.md) (2026-10-24): A fair checklist for teachers and editors who suspect AI: what a detector score can prove, what it cannot, and how to talk to the author before deciding.
- [AI detectors for students: avoiding false accusations](https://authortext.com/en/articles/students-false-accusation.md) (2026-10-24): What an AI detector score means for a student, how to keep proof of your own work, and what to do if a teacher flags your essay by mistake.
- [Burstiness and perplexity without the maths](https://authortext.com/en/articles/burstiness-perplexity.md) (2026-10-24): Burstiness and perplexity explained in plain words: what they mean for a reader, how they relate to our signals, and why they are weak evidence alone.
- [Checking a freelancer's text: a workflow for clients](https://authortext.com/en/articles/checking-freelancer-text.md) (2026-10-24): A practical workflow for clients who hire writers: a clear brief, drafts, style consistency, fair use of a checker and sensible AI wording in a contract.
- [How to read the numbers in an AuthorText analysis](https://authortext.com/en/articles/reading-the-numbers.md) (2026-10-24): What sentence-length variation, repetition, stock phrases and connective density mean in an AuthorText analysis, with a tiny example and honest limits.
- [Invisible characters in copied text](https://authortext.com/en/articles/invisible-characters.md) (2026-10-24): Zero-width spaces, soft hyphens, curly quotes and long dashes: where they come from in copied text, what they break, and how to clean them safely.
- [What is new in AI detection: verified or not](https://authortext.com/en/articles/detection-news-2026.md) (2026-10-12): What is new in 2026: benchmarks with mixed authorship, surprisal-rhythm features, EU marking rules. What we verified and what we did not.
- [How AI text detection works today](https://authortext.com/en/articles/how-detection-works.md) (2026-10-11): Seven ways to check a text for AI, what limits each, and why an honest result is an estimate with caveats, not a percentage of AI authorship.
- [Why modern AI detectors get it wrong](https://authortext.com/en/articles/why-detectors-err.md) (2026-10-08): A detector scores style, not the author. Why polished human writing gets high scores and how to read a result sensibly.
- [When the detector mistakes a human for AI](https://authortext.com/en/articles/when-detectors-disagree.md) (2026-10-07): An article written by a person scored "AI" while a machine-written text looked "more human". Why that happens and what to do about authorship checks.
- [How to make your text sound more alive](https://authortext.com/en/articles/make-text-livelier.md) (2026-10-06): Six techniques for livelier writing: concrete detail, uneven rhythm, fewer stock phrases, your own voice. No promises to "beat the detector".
- [How to tell if a text was written by AI](https://authortext.com/en/articles/signs-of-ai-text.md) (2026-10-06): Four signs of AI-written text: an even rhythm, stock phrases, general words with no specifics. And why none of them is proof.