The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Short answer: The Wall Street Journal reported in August 2024 that OpenAI had developed a text-watermarking system for identifying likely ChatGPT-generated writing. The paper described an internal OpenAI document claiming 99.9% certainty. That figure has not been published as an independently reproducible accuracy benchmark, and OpenAI had not released the watermark in the public status information available from August 2024.
What OpenAI reportedly built
The reported system is a text watermark, not a conventional detector that simply learns stylistic patterns. OpenAI says the method subtly influences how a model selects successive words or tokens. Those choices create a statistical pattern that a separate test can look for later.
As an Amazon Associate I earn from qualifying purchases.
The result would indicate the probability that a passage came from ChatGPT. It would not prove who typed, edited or submitted the text, and the Wall Street Journal described the reported watermark as applying to ChatGPT output rather than writing from competing AI providers.
What “99.9% certainty” means
The 99.9% certainty figure comes from a Wall Street Journal report in 2024 describing an internal OpenAI document. Public materials do not disclose the test corpus, decision threshold, minimum passage length, language coverage, false-positive rate or independent replication for that method. It should therefore be reported as “99.9% certainty — The Wall Street Journal, 2024, describing an internal document,” not as a universally verified 99.9% accuracy rate for all AI writing.
#1 Best Overall
- Upgraded AI-Powered Detection: Military-grade technology detects hidden cameras, listening devices, and GPS trackers with precision. Enjoy peace of mind in hotels, offices, and even your own home. Stay one step ahead of hidden threats!
- Simple, Fast & Effective: Just turn it on, sweep the area, and let the audible alarm + LED alerts notify you of threats. No technical skills needed - Press, Search, Relax! Skip expensive private investigators - protect yourself in seconds.
- Compact & Travel-Ready: Lightweight, rechargeable, and pocket-sized for discreet, on-the-go security. Toss it in your bag, purse, or pocket - perfect for travel, work, and public spaces.
- Total Privacy Protection: Don’t gamble with your security. Safeguard against spying in hotel rooms, changing rooms, offices, cars, dorms, and more. Know for sure if you’re being watched, recorded, or tracked.
- Trusted by Experts & Customers: Designed with cybersecurity and counter-surveillance professionals. Join 300,000+ satisfied users who rely on our detectors for ultimate privacy & safety.
Why this is different from OpenAI’s old AI Classifier
OpenAI’s later watermark is often confused with its 2023 AI Classifier. They are separate systems with different evidence and failure modes.
| Feature | 2023 AI Classifier | Reported text watermark |
|---|---|---|
| Signal | A classifier’s judgment based on text patterns | An embedded statistical pattern created during token selection |
| Scope | Attempted to classify AI-written text from multiple systems | Reportedly targeted text generated by ChatGPT |
| Public status | Released experimentally, then discontinued on July 20, 2023 | Reported as unreleased; OpenAI’s August 4, 2024 update said the company was still considering it |
| Published performance | 26% true-positive rate on AI text and 9% false-positive rate on human text in OpenAI’s English challenge set | 99.9% certainty reported by the Wall Street Journal from an internal document; matched public test details were not provided |
| Known weaknesses | Very unreliable below 1,000 characters; weaker on code and languages other than English; vulnerable to edits and calibration errors | OpenAI said global rewriting, translation and retranslation, or rewriting with another generative model could make circumvention trivial |
Because the two systems were not tested on the same data or under the same definitions, their percentages cannot be compared as if they were competing benchmark scores.
Why OpenAI did not release the reported watermark
OpenAI’s August 2024 public update said the watermark was technically promising but that the company was researching alternatives and weighing important risks. The Wall Street Journal reported that the project had been debated internally for about two years and had been ready for release for roughly a year.
Recommended Free Tools
Circumvention is easy after substantial rewriting
OpenAI said localized tampering may leave the watermark detectable, but broader transformations can defeat it. The company specifically cited global paraphrasing, translation followed by retranslation, rewriting with another generative model, and inserting or removing a character between words. That means a detector could work well on untouched ChatGPT output yet provide little evidence about a passage that has been substantially revised.
Risk of disproportionate effects
OpenAI also warned that a watermark could stigmatize people who use AI as a writing aid, including non-native English speakers. The company’s concern was not limited to cheating: a tool that labels text as AI-generated can affect students, job applicants, researchers and anyone whose writing is concise or formulaic.
The Wall Street Journal said an OpenAI survey found that nearly one-third of loyal ChatGPT users would be turned off by anti-cheating technology. That is an account of the company’s survey, not a population-wide estimate.
What the discontinued classifier showed about false accusations
OpenAI described its 2023 classifier as “not fully reliable” and said it should not be the primary basis for decisions. In its English challenge set, it correctly identified only 26% of AI-written text as likely AI-written while falsely labeling 9% of human-written text as AI-generated. OpenAI said performance improved with longer passages but was very unreliable below 1,000 characters.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallRank #2
- AI-Powered Smart Signal Analysis – This camera detector, built in intelligent AI chip processes RF signals in real time, reducing background interference and false alarms for more reliable detection of hidden wireless cameras, audio bugs, and GPS trackers. The adjustable sensitivity dial lets you fine-tune scanning levels to match different environments – from busy hotels to quiet homes – so you get precise alerts without constant beeping.
- 4-in-1 Detection with 5-Level Sensitivity – As a reliable hidden camera detectors, it scans for wireless cameras, hidden pinhole lenses, GPS trackers, and magnetic field devices, giving you complete coverage against various surveillance threats. The 5-level adjustable sensitivity lets you dial in the right detection range for any setting – turn it up for weak signals in large spaces, or lower it in crowded areas to reduce interference. Works great at home, in hotel rooms, and while traveling, so you always know your privacy is protected.
- Easy Operation with 4 Modes – With a built-in GPS tracker detector, switch between wireless signal scanning, hidden camera finder, magnetic field detection, and flashlight – all in one compact device. Select between sound or vibration alerts for quiet, discreet scanning in any environment.
- Long Battery Life & Portable Design – Built-in 800mAh rechargeable battery provides up to 25 hours of continuous use. Fully charges in just 1.5 hours via USB. Small enough to carry anywhere – weighs next to nothing, so you can take it on every trip.
- OpenAI recommended English-only use for that classifier.
- Code and other languages performed worse.
- Predictable writing could not be reliably attributed to a machine.
- Editing could evade detection, and confident false labels could occur.
- Calibration could fail when text differed from the data used to train the classifier.
Those figures describe the withdrawn classifier, not the later watermark. They nonetheless demonstrate why a detector score should be treated as a lead for review rather than proof of misconduct or authorship.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Evidence that detector errors can be unfair
OpenAI’s educator guidance reported that its detector labeled human-written works, including Shakespeare and the Declaration of Independence, as AI-generated. It also noted possible disproportionate effects on students learning English and on people whose writing is formulaic or concise.
A 2023 study by Weixin Liang, Mert Yuksekgonul, Yining Mao, Eric Wu and James Zou evaluated seven detectors on 91 TOEFL essays by non-native English writers and 88 U.S. eighth-grade essays. Across the seven tools, the average false-positive rate on the TOEFL sample was 61.22%; the authors reported near-perfect accuracy on the eighth-grade sample. These results are specific to those detectors and datasets. They do not measure OpenAI’s later watermark and should not be generalized to every detector.
Can ChatGPT tell whether it wrote a passage?
No. OpenAI’s educator FAQ says ChatGPT has no “knowledge” of what content could be AI-generated or what it generated. Asking ChatGPT to identify its own text can produce an invented or overconfident answer, so that response is not an authorship test.
Free tools Windows power users keep installed
One-click scans. No signup required.
What schools and employers should do with detector results
A watermark or classifier result can be one signal in a broader inquiry, but it should not be the sole basis for discipline, rejection or an accusation. A fair review should consider the assignment or document history, drafts, notes, version records, the writer’s explanation and whether the alleged tool has published validation for the relevant language and text length.
For the reported watermark specifically, the public record available from OpenAI’s August 4, 2024 update establishes development and continued evaluation, not a confirmed public launch. The Wall Street Journal’s 99.9% figure remains an attributed internal claim rather than an independently verified universal accuracy rate.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
