AI
“Artificial intelligence detectors are facing major scrutiny after popular software flagged the 1776 US Declaration of Independence as 99% AI-generated." (Post carries a red "BREAKING" banner; image inset shows "99.99% AI GPT*")”
Plain restatementAt least one commercial AI-text-detection tool returned a result of approximately 98 to 99.99 percent "AI-generated" when given text from the 1776 US Declaration of Independence, and this is cited as evidence that AI detectors produce false positives on formal or historical human writing.
Distortion code this site does not recognise yet: rumor_as_fact. Not collectible until the field guide has an entry.
This one is real in substance. AI text detectors have repeatedly flagged the 1776 Declaration of Independence as almost entirely AI-written, and journalists and data scientists have reproduced the result several times, getting scores of about 97 to 98 percent between 2024 and 2025. One tool, ZeroGPT, is responsible for most of the documented failures, and the same tool also flagged the 1836 Texas Declaration of Independence at 86 percent. The explanation in the post is broadly right: famous documents appear constantly in AI training data, which makes their wording look statistically unsurprising to a detector. Three things in the post go too far. It is labeled BREAKING, but this story dates to July 2023 and the specific viral screenshot is from November 2025. It says "AI detectors" generally, when in the best documented test three of four detectors correctly called the text human. And the headline 99.99 percent figure comes from a single social media screenshot that nobody has reproduced, while the verified numbers are lower. The bottom line the post draws is still sound and is backed by peer-reviewed research: these tools produce false positives often enough that they should not be treated as proof of anything on their own.
[drifted from the evidence:] Artificial intelligence detectors are facing major scrutiny after popular software flagged the 1776 US Declaration of Independence as [drifted from the evidence:] 99% AI-generated." (Post carries a red "BREAKING" banner; image inset shows "99.99% AI [drifted from the evidence:] GPT*")
[added by the neutral restatement:] At least one commercial AI-text-detection tool returned a result of approximately 98 to 99.99 percent "AI-generated" when given text from the 1776 US Declaration of Independence, [added by the neutral restatement:] and this is cited as [added by the neutral restatement:] evidence that AI [added by the neutral restatement:] detectors produce false positives on formal or historical human writing.
Red-tinted words in the claim drifted from the evidence. Green-tinted words are what a neutral restatement needs.
The trace / claim to source
- An AI detector did return a near-certain "AI-generated" verdict on the human-written 1776 Declaration of Independence. This is not a hoax and not a one-off.
- The result has been independently reproduced by multiple named parties on the same tool, with scores of 97.75%, 97.93%, and 98.51% across 2024 and 2025, and on a different founding document at 86.54%.
- The mechanism the caption gives is directionally correct as applied to perplexity-based detectors, and matches what the detector industry itself said in 2023: heavily reproduced canonical texts sit in model training data, produce very low perplexity, and therefore score as machine-like.
- The caption's second mechanism, that these documents are in LLM training corpora, is the explanation the founder of GPTZero gave on the record to Ars Technica.
- The general reliability warning is supported by refereed research, not just viral anecdote: a peer-reviewed Patterns paper measured a 61.22% average false positive rate on non-native English essays across seven detectors.
- The conclusion that detectors should not be treated as definitive proof is supported by the vendors' own conduct: OpenAI withdrew its classifier for low accuracy, and Turnitin revised its false positive claim upward and then stopped publishing it.
- Date context mismatch: the post uses a red "BREAKING" banner and says detectors are "facing major scrutiny after" this result. The Constitution version of this story was reported by Ars Technica in July 2023, the Declaration version by Decrypt in October 2024, and the Texas version in May 2025. The specific 99.99% screenshot circulated in November 2025, roughly ten months before this post. Nothing here is breaking, and the scrutiny predates the incident being used to justify it.
- Subgroup generalization: the caption says "AI detectors" and "detection software routinely misinterprets classic human writing." In the best-documented test of this exact text, three of four detectors got it right. Quillbot called it 100% human, GPTZero called it 89% likely human, and Grammarly performed best. One tool failed. Presenting one tool's failure as the behavior of the category overstates what the evidence shows, even though the category does have a documented false positive problem for other reasons.
- Omitted qualifier: the post never names the detector. The failure is specific and attributable, and naming it is the single most useful piece of information for a reader deciding whether to trust a tool. Leaving it out converts an actionable finding into a vague indictment.
- Rumor as fact: the headline number 99.99% comes from one consumer screenshot posted to Reddit, with no stated input passage, no date of run, and no reproduction. The figures that have actually been reproduced by people who published their method are 97.75 to 98.51 percent. The post presents the highest and least verified number as the finding.
- Date context mismatch (second instance): the caption states that "most AI detectors analyze text based on perplexity and burstiness." GPTZero's own support documentation states that as of autumn 2023 it no longer uses perplexity and burstiness, having migrated to a deep-learning architecture. A 2023-era description of how detectors work is presented as current practice in 2026.
- The caption defines burstiness as "variations in sentence length." GPTZero's own documentation defines it as how much perplexity varies over the document. These are related but not the same thing, and the post states the looser version as technical fact.
- The identity of the detector in this specific post's screenshot. The visible string "99.99% AI GPT*" matches ZeroGPT's output format, and a social post from the same week names ZeroGPT, but a social post is not adequate attribution on its own. I could not retrieve the original Reddit thread.
- Whether the 99.99% result is currently reproducible. I did not run any detector, and these tools update silently without version numbers. A result from November 2025 does not establish behavior on 2026-09-22.
- What text was submitted. "The Declaration of Independence" could mean the preamble, the full document, or the grievances section, and detector output is length-sensitive. No coverage I found specifies this for the viral screenshot.
- Whether the post's image itself is an authentic screenshot or a recreation. I have only the intake description of the image, not the file's provenance, so I make no finding on the artifact.
- ZeroGPT's marketed 98.5% accuracy figure is a vendor self-report with no methodology published. Third-party accuracy estimates I encountered were mostly on commercial review and competitor sites with their own interest in the answer, so I do not treat any specific counter-figure as established.
The underlying phenomenon is real, repeatedly reproduced, and independently documented by multiple parties who ran the test themselves. Decrypt ran the Declaration's preamble through four detectors in October 2024 and reported that ZeroGPT said the Declaration of Independence text was 97.93% AI-generated, while Quillbot identified the same text as "Human-written 100%" and GPTZero gave it an 89% probability of being written by humans. Decrypt described using the same excerpt that data scientist Christopher Penn had used, and separately testing an excerpt from E.M. Forster's 1909 story "The Machine Stops" rewritten by ChatGPT. In that second test, ZeroGPT identified the genuine human "The Machine Stops" text as likely human at 4.27%, but also mistakenly labeled the AI-generated version as human-written at 6.35%, meaning the tool failed in both directions in the same sitting. Independent repeats produced similar numbers. The Sentinel reported that Zero GPT determined the document was 97.75% AI-generated, while Copy Leaks and Writers.com said it was human-written, and noted that Zero GPT is the first link that appears when searching for AI detectors on Google. The Dallas Express extended the test to a different founding document: ZeroGPT reported that the 1836 Texas Declaration of Independence was 86.54% AI GPT, despite the document being signed by 59 delegates at Washington-on-the-Brazos in March 1836, nearly 190 years before the rise of ChatGPT. That outlet also recorded the earlier result and the reaction it drew: Decrypt's October 2024 finding of 97.93% prompted Christopher Penn, chief data scientist at Trust Insights, to call AI detectors a "joke," warning they were "unsophisticated and harmful," and to say that "these tools are being used to do things like disqualify students, putting them on academic probation or suspension". The specific 99.99% figure in this post is newer and thinner. It traces to a screenshot circulated in late November 2025: a Reddit user uploaded the 1776 text into an AI-detection tool and the detector declared the document "99.99 per cent AI-generated," after which screenshots spread quickly across social media. A social post from the same week names the tool as ZeroGPT. I could not retrieve the original Reddit thread, and no outlet I found reproduced the run under stated conditions. The mechanism the caption describes is broadly supported by the original reporting on the earlier Constitution version of this story. Ars Technica's July 2023 piece explained that if the language in a piece of text is not surprising based on the model's training, the perplexity will be low, so the detector is more likely to classify it as AI-generated, and the Constitution's language is so ingrained in these models that they classify it as AI-generated, creating a false positive. GPTZero's founder told that outlet that "The US Constitution is a text fed repeatedly into the training data of many large language models". The broader reliability critique is supported by refereed work, not just anecdotes. Liang et al. found that seven detectors misclassified over half of a set of TOEFL essays as "AI-generated," with an average false positive rate of 61.22%, with all seven unanimously identifying 18 of the 91 essays as AI-authored. The published version states that the authors "exposed an alarming bias in GPT detectors against non-native English speakers: over half of the non-native English writing samples were misclassified as AI generated, while the accuracy for native samples remained near perfect". Separately, a 2023 study by Weber-Wulff et al. evaluated 14 detection tools, and its authors reported accuracy in many studies of "only around 50% or slightly above". Vendor conduct corroborates the limitation. OpenAI's own announcement page now carries the note that "As of July 20, 2023, the AI classifier is no longer available due to its low rate of accuracy". Turnitin's chief product officer, after launching with a sub-1% claim, conceded a higher figure: the tool has a higher false positive rate than the company originally asserted, and the company has not disclosed the new document-level false positive rate. Finally, the tool most often implicated markets itself confidently while disclaiming in the same breath. Its product page advertises 98.5% accuracy for sentence-level detection, while also stating that "No AI detector is 100% accurate, machine learning predictions are probabilistic by nature".
Complete reasoning
The reply is formatted for pasting into the thread where the claim is circulating.
Compact share page: ai.trueseeker.com/s/279cfda7b852/As8O0vvcUEK5WoW8fmt34Jr5XUE
Ask this case
Answers come only from the case file above; nothing is added.
Did an AI detector really score the Declaration of Independence as almost entirely AI-generated?
Yes. Multiple people independently ran the text through detection tools and got very high AI scores, including 97.75 percent, 97.93 percent, and a later screenshot claiming 99.99 percent, mostly from a tool called ZeroGPT.
Is this a new discovery, as the BREAKING banner suggests?
No. The underlying story goes back to July 2023 when Ars Technica reported a similar false positive on the US Constitution. The Declaration version was reported in October 2024, and the viral 99.99 percent screenshot is from November 2025.
Do all AI detectors flag the Declaration this way?
No. In the most documented test, three of four detectors correctly identified the text as human-written, and only ZeroGPT returned a high AI score. The post's claim that 'AI detectors' generally failed overstates what happened.
Why would a detector think a 1776 document is AI-generated?
Detectors that rely on perplexity flag text as AI-made when its wording is statistically unsurprising to the model. Famous documents like the Declaration appear repeatedly in AI training data, so their language reads as predictable, which triggers a false positive. GPTZero's founder gave this same explanation on the record.
Should this result be taken as proof that AI detectors are unreliable in general?
The case file supports that conclusion, though not from this one example alone. A peer reviewed study found detectors misclassified over half of a set of non native English essays as AI generated, and OpenAI withdrew its own detector in 2023 citing low accuracy.