Is Jane Eyre AI-generated? 2 of 7 AI detectors said yes
No. Of the 7 AI detectors we tested, only 2 flagged the opening of Charlotte Brontë's 1847 novel, Fakespot with a 96.2% AI probability and TMR with 97.6%. Both flag a large share of human writing, 34.9% and 83.3% of the human texts in our main test. Desklib gave the passage a 0.0% AI probability.
Tested
What each AI detector said
| Detector | Verdict | AI probability |
|---|---|---|
| Fast-DetectGPT (Qwen2.5-1.5B) | Human | raw score only |
| Desklib AI text detector | Human | 0.0% |
| Fakespot AI text detector | AI-generated | 96.2% |
| MAGE | Human | raw score only |
| E5-small LoRA detector | Human | 16.3% |
| OpenAI RoBERTa detector | Human | 0.0% |
| TMR (Target Mining RoBERTa) | AI-generated | 97.6% |
The passage we tested
There was no possibility of taking a walk that day. We had been wandering, indeed, in the leafless shrubbery an hour in the morning; but since dinner (Mrs. Reed, when there was no company, dined early) the cold winter wind had brought with it clouds so sombre, and a rain so penetrating, that further outdoor exercise was now out of...
One confident detector against the rest
Fakespot was close to certain, at 96.2%, and TMR agreed at 97.6%. TMR’s verdict can be set aside first, because it flagged every human fiction passage in our main test and responds mainly to paragraph breaks. That leaves Fakespot on its own against Desklib (0.0%), the E5-small LoRA detector (16.3%), the OpenAI RoBERTa detector (0.0%), MAGE and Fast-DetectGPT.
Fakespot flagged 34.9% of all human texts in our main test but only 12.5% of human fiction, so a flag on a novel is less common for it. On this passage it still produced one of its highest scores of the 30 famous texts.
Polished sentences and a famous first line
Jane’s narration is formal and carefully balanced. Sentences run on through semicolons and colons (“I was glad of it: I never liked long walks, especially on chilly afternoons: dreadful to me was the coming home in the raw twilight…”), and the vocabulary leans Latinate: “dispensed from joining the group”, “endeavouring in good earnest to acquire a more sociable and childlike disposition”. Mrs. Reed’s reported speech is one long, polite, perfectly structured sentence. Polished, grammatical prose with no slips is one thing a classifier may associate with AI text, and that is a plausible reason for Fakespot’s flag.
The other likely explanation, that a language model has memorized the text, fits less well here. “There was no possibility of taking a walk that day” is a famous first line, but Fast-DetectGPT, the detector that reacts to familiar wording, passed the passage. Brontë’s later sentences, such as “I was shrined in double retirement” behind “the red moreen curtain”, are probably too unusual for a model to predict well.
When one detector is very sure
Fakespot’s 96.2% looked decisive and was wrong. If you write in a careful, formal style, with long sentences and few contractions, it may flag your work too. Little Women drew a similar Fakespot score. The false positives page shows how often the same thing happens to ordinary human writing. Every other classic we ran is in the famous texts overview.
Questions people ask
Why did Fakespot say Jane Eyre is AI-generated?
The model gives no reasons, but a likely factor is the register, long formal sentences full of semicolons and Latinate words, which may resemble the polished prose of AI text more than casual human writing does.
Did Fast-DetectGPT flag Jane Eyre?
No. Fast-DetectGPT, which flags text a language model finds predictable, passed this passage, and it did not flag a single human fiction passage in our main test.
How sure was Fakespot?
It gave the passage a 96.2% AI probability. A high score from one detector is still weak evidence, since Fakespot flagged 34.9% of human texts in our main test.