What this means
This text shows several patterns common in AI-generated writing — uniform sentence lengths and predictable phrasing. That is a signal, not proof: formulaic, heavily edited, or very formal human writing can score high too.
Suggested next steps
- Wrote this yourself? High scores do happen to human writing — see why detectors give false positives. Drafts and version history are stronger evidence of authorship than any score.
- Reviewing someone else's work? Treat this as one signal to look closer, not a conclusion — see how to use detectors responsibly.
- Short passages make every detector less reliable — try a longer sample and compare. More in how accurate are AI detectors.
What this means
This text mixes AI-like patterns with human-like variation. That is common for AI-assisted drafts and edited AI output — and also for ordinary human writing in a formal, structured style. This range is genuinely ambiguous.
Suggested next steps
- Avoid drawing conclusions from a mid-range score alone — it is the least informative band. See what detector scores really mean.
- Reviewing someone's work? A conversation about how it was written tells you more than this number — see how to use detectors responsibly.
- Longer, more varied samples give the signals more to work with.
What this means
This text shows the varied sentence rhythm and vocabulary typical of human writing. A low score is still not a certificate: carefully edited AI text can read as human, and no detector can rule AI involvement in or out.
Suggested next steps
- Treat this as "no strong AI signal," not proof of human authorship — see how accurate are AI detectors.
- Reviewing someone's work? A low score is not proof either way — the same care applies as with a high one: how to use detectors responsibly.
Sentence Analysis (red = likely AI · green = likely human)
How Does AI Detection Work?
AI-generated text tends to show measurable statistical patterns that differ from human writing. This detector computes five of them, entirely in your browser, and combines them into a weighted score:
- Sentence uniformity — the standard deviation of sentence length relative to the average. Human writing varies far more than AI writing does.
- Burstiness — whether long and short sentences cluster, or arrive at an even pace. Measured here from the same spread of sentence lengths.
- Vocabulary diversity — the ratio of distinct longer words to total words. Human writers reach for a wider range.
- Transition-word density — how many stock connectors like "furthermore," "moreover" and "in conclusion" appear relative to sentence count.
- Sentence-opening repetition — how often sentences begin the same way. AI output repeats openings more than people do.
What this is not. Some detectors use a language model to compute true perplexity — how surprised a model is by your next word. This one does not, and cannot: it is a few kilobytes of JavaScript running on your device with no model behind it. What it measures is structural regularity, which correlates with AI authorship but is not the same thing. We would rather tell you exactly what runs than borrow the vocabulary of a heavier method.
Is This Detector 100% Accurate?
No AI detector is 100% accurate — including this one. False positives (flagging human text as AI) and false negatives (missing AI text) are possible. This tool is best used as one signal among several, never as proof.
Three specific limits worth knowing, because they follow directly from how it works:
- The score never leaves the 5–95 range. It is deliberately incapable of saying "definitely AI" or "definitely human", because the underlying signals do not support that certainty.
- Transition words are counted once each. Using "furthermore" twenty times registers the same as using it once, so the signal detects variety of stock phrasing rather than volume.
- Short text is unreliable. Every signal is statistical, and statistics on four sentences are noise. Paste several paragraphs for a result worth anything.
Never use this as the sole basis for an accusation. Formal, non-native and heavily edited writing all trigger the same structural regularity that AI output does. Read why AI detectors give false positives before you act on any score.
Who Uses AI Detectors?
- Teachers & professors — checking student submissions
- Publishers & editors — ensuring editorial standards
- Content managers — verifying freelancer work
- Hiring managers — reviewing job applications
- Researchers — maintaining academic integrity
Learn More About AI Detection
- How accurate are AI detectors? — what detector scores really mean, and why no tool can promise certainty
- Why AI detectors give false positives — why human writing gets flagged, and what to do if yours is
- How to use AI detectors responsibly — a practical guide for teachers, editors, and managers