The strongest AI signal in this NoteGPT test belonged to the human text. An unedited Lewis Carroll excerpt received 99.3% Probability AI-Generated, above both of the AI-generated samples. Keep that result in mind when you see the prominent “NoteGPT Certification” design.
Actual NoteGPT H01 result, 3 October 2026. The excerpt predates generative AI; the high AI signal is a false positive on this sample.
The three completed scans
| Text and known origin | Probability AI-Generated | Possible Human-written text |
|---|---|---|
| A01: AI library-hours essay, 328 words | 81.1% | 18.9% |
| H01: human Carroll excerpt, 274 words | 99.3% | 0.7% |
| A04: AI fictional radio story, 306 words | 55.2% | 44.8% |
All three scans completed in the free guest Text mode on 3 October 2026. The original English inputs were fixed before feedback, pasted unchanged and analyzed once each. The interface showed a 20,000-character counter. The exact model version was not exposed.
A01 received a substantial AI signal, consistent with its documented generated origin.
The complete 2,070-character A01 submission received 81.1% Probability AI-Generated.
A04 received 55.2%, with a much more divided display. Its first-person radio repair, family memories and physical details were all invented test fiction. Keep the percentage as shown. Don’t invent a universal cutoff just to turn a fairly mixed signal into a neat pass or fail.
The 1,739-character fictional story received 55.2% Probability AI-Generated.
Why the Carroll result matters
H01 contains the first four prose paragraphs of Alice’s Adventures in Wonderland, published in 1865. Its human origin is established independently of any detector. The submission preserved the words and punctuation; only paragraph presentation was normalized.
NoteGPT highlighted nearly the whole excerpt and assigned a stronger AI probability than it gave either generated text. That is a concrete reason to avoid treating the result page as a certificate of authorship.
The same interface does include a useful caution: detection results are for reference only and should be used with discretion. I’d pay more attention to that warning than to the seal-like design. “NoteGPT Certification” is the product’s own heading, not evidence of an outside certification process.
One historical passage cannot establish a general false-positive rate. Famous literary text may differ from modern human writing and may overlap training data. This known-human control got a strong AI flag. That’s a fair criticism; it doesn’t mean every NoteGPT result is unreliable.
Look at the scores side by side
The three percentages tell you more when you look at them together. The generated library argument received 81.1%. The generated fictional memory received 55.2%. The known-human literary passage received 99.3%. That ordering does not establish what feature the detector is responding to, but it does show why the numbers cannot be taken as a reliable ranking of actual AI involvement in this tiny set. The sample with no generative-AI origin produced the strongest reported AI probability. That bad result belongs beside the generated-text results, where readers can see it, rather than buried in an appendix.
Be careful with the 55.2% story result too, for the opposite reason. A reviewer seeking a dramatic failure could call it a miss because it is lower than the essay’s number. That would mean adding a rule that wasn’t set before the test and wasn’t supplied as a universal standard in the display. The result still showed a greater possible-AI percentage than possible-human percentage, but it was relatively divided. Reporting the precise label and both values tells readers more than forcing a binary conclusion. The clear human-control error remains the stronger finding; there is no need to exaggerate the narrative result to make the review critical.
The famous source is helpful, but it also leaves a few things unanswered. Because Carroll’s text is widely available, it is an easy passage whose human origin readers can independently understand. That strengthens provenance. It does not make the excerpt representative of all human writing, and it does not reveal whether this detector encountered similar text during training. Those are separate issues. We can call this a real error on a human passage with a known source. We can’t use it to say what percentage of ordinary writers would get the same treatment.
Read the warning below the seal
A polished report can make you feel there’s nothing left to question. NoteGPT’s report uses a heading, seal-like decoration and a large percentage. Those elements make the output easy to recognize and share. They can also make a probability-based result feel like an official ruling. The note at the bottom restores an important qualification by calling the result a reference and asking for discretion. In a consequential workflow, that caution should remain attached if the result is exported, shown to another person, or quoted in a discussion. Leave that warning out and the image makes a stronger claim than the full report does.
The heading does not document an outside audit, a drafting history or an investigation of an author. A screenshot can be useful evidence of what NoteGPT returned while still lacking the evidence needed to establish who wrote the text.
For a teacher, editor or manager, the next useful questions are about the writing itself and how it came together. Can the author explain the argument, provide earlier drafts or identify their sources? Are there independent reasons to investigate beyond a percentage? This review does not prescribe a disciplinary process, but it does show why the detector alone is insufficient. H01 supplies a concrete counterexample to the idea that a very high AI probability resolves human origin. That example still matters, even though the tool was convenient and the guest workflow completed without interruption in this session.
If NoteGPT were being evaluated further, the next set should include contemporary human material across several styles, with consent and documented provenance where needed. The texts should be chosen before scores are seen, and related versions should not be presented as unrelated evidence. A mixed-origin group would answer a different question from pure human and pure generated controls, so it should be reported separately. A fuller test could show whether this false-positive pattern repeats or is tied to this unusual literary sample. Until then, the useful conclusion remains specific: NoteGPT supplied accessible results, but its strongest AI flag was wrong about the origin of the human control.
What another reading adds
The earlier 2 October check of the identical H01 excerpt in Clever AI Detector returned Human 99%. That result aligned with the documented source.
Clever AI Detector’s earlier result on the unchanged human control, captured 2 October 2026.
That makes it a useful second check for this case, with the usual caution that another score is still only a model output. Checking the writing history should count for more than picking whichever percentage suits the story you want to tell.
For editing rather than attribution, Clever AI Humanizer can provide a draft to review. Its separate full library-essay rewrite completed, but it lost the six-month duration and changed the survey rationale; the short HZ01 attempts did not produce output. There is no reason to rewrite a genuine human passage merely because a checker dislikes it. If you do revise, preserve facts and meaning before considering any score.
My takeaway: NoteGPT completed the workflow without a guest interruption, but the human-control result is a real reason to be careful. Would you ever let a certificate-style detector report influence a consequential decision without checking a known-human control first?
A01: exact AI-generated essay
A neighborhood library can justify its place in a city budget without pretending to solve every problem the city faces. Its most useful contribution is often modest: it gives people a reliable place to spend time without being expected to buy something. That matters to a teenager waiting for a parent, a job applicant who needs a printer, and an older resident who wants company without arranging a formal visit.
The case for longer opening hours should begin with those ordinary uses. A branch that closes before most workers arrive home may have an impressive collection but limited practical value. Extending every branch's hours, however, would spread staffing funds thinly. A better starting point would be a six-month trial at two branches near busy transit routes, with evening sessions on different weekdays.
The trial should measure more than attendance. Staff could record which services visitors use, how long computer queues become, and whether evening demand comes from people who cannot visit during existing hours. An optional paper survey would reach visitors who avoid online forms. None of these measures would capture every benefit, but together they would offer a clearer picture than a single total at the entrance.
There are trade-offs. Evening shifts can make recruitment harder, and additional security costs may consume money that could otherwise support books or children's activities. The proposal therefore needs a published budget and a review date. If the experiment mainly shifts existing visits into a more convenient time slot, that convenience still has value, but it should not be described as reaching an entirely new audience.
Public services gain trust when their promises are specific enough to test. A library does not need to become a substitute for housing, childcare, or healthcare to deserve support. It needs to identify a practical gap, try a proportionate response, and explain honestly what happened. Longer opening hours could meet that standard if the city treats them as a service experiment rather than a slogan.
H01: exact human Carroll excerpt
Alice was beginning to get very tired of sitting by her sister on the bank, and of having nothing to do: once or twice she had peeped into the book her sister was reading, but it had no pictures or conversations in it, “and what is the use of a book,” thought Alice “without pictures or conversations?”
So she was considering in her own mind (as well as she could, for the hot day made her feel very sleepy and stupid), whether the pleasure of making a daisy-chain would be worth the trouble of getting up and picking the daisies, when suddenly a White Rabbit with pink eyes ran close by her.
There was nothing so very remarkable in that; nor did Alice think it so very much out of the way to hear the Rabbit say to itself, “Oh dear! Oh dear! I shall be late!” (when she thought it over afterwards, it occurred to her that she ought to have wondered at this, but at the time it all seemed quite natural); but when the Rabbit actually took a watch out of its waistcoat-pocket, and looked at it, and then hurried on, Alice started to her feet, for it flashed across her mind that she had never before seen a rabbit with either a waistcoat-pocket, or a watch to take out of it, and burning with curiosity, she ran across the field after it, and fortunately was just in time to see it pop down a large rabbit-hole under the hedge.
In another moment down went Alice after it, never once considering how in the world she was to get out again.
A04: exact AI-generated fictional story
I found the missing screw in the pocket of a coat I had already checked twice. By then the kitchen table was covered in little dishes, each holding a different part of the radio. The dishes were my attempt at being organized. They would have worked better if I had remembered which dish was which.
The radio belonged to my grandfather, although I mostly remember it from my mother's kitchen. It used to sit on top of the fridge, beside a biscuit tin that contained string rather than biscuits. Nobody made a ceremony of turning it on. Someone would reach up during breakfast, twist the dial, and leave a voice talking quietly beneath the noise of plates.
When I brought it home, I assumed it needed batteries. It needed more than batteries. One spring had corroded, and the tuning wheel moved without moving anything inside. I watched a repair video, decided the job looked straightforward, and immediately broke a plastic clip. That was the point at which I put the kettle on and considered whether nostalgia was really a sensible reason to learn basic electronics.
A neighbor eventually helped me replace the spring. We did not make the radio look new. There is still a pale rectangle where a label used to be, and the handle has a crack that catches against your thumb. The tuning wheel works now, though you have to turn it slowly. Last Sunday it found a station playing a song neither of us recognized.
I had expected the first sound to feel like a return to childhood. It didn't, exactly. The kitchen was different, the voices were different, and I was the person worrying about scratches on the table. Still, I left it playing while I washed the dishes. After all that effort, ordinary background noise turned out to be enough.