AI Writing Check (AIWritingCheck.org)

One record in the studies index tested the tool. Liang 2023 accessed seven detectors on March 15, 2023, listing the second as "Quil.org (aiwritingcheck.org)", and ran 91 human-written TOEFL essays and 88 US eighth-grade essays through all of them. Across the seven it reports an average false positive rate of 61.22% on the TOEFL essays and 5.19% on the US essays; 18 of the 91 TOEFL essays (19.78%) were flagged by all seven detectors at once and 89 of 91 (97.80%) by at least one; after a prompt asking ChatGPT to enhance the word choices, the TOEFL average fell to 11.77% (arXiv 2304.02819v3, read 8 November 2026). Per-detector rates appear only as a bar chart and are not printed as numbers, so the card carries the averages with their bases, as every card tested in Liang 2023 does.

The fields follow the index rules. The false positive value stays empty because no count is attributed to the tool, and the vendor's "one or two of every ten" is a description of its own estimate, not a measurement. The record counts as the card's one non-English record, because the 91 TOEFL essays were written in English by non-native writers, the class the languages evidence ranking tracks. No record attacked the tool, so the robustness field is empty, though Liang 2023's self-edit prompt cut detection on 31 college essays from 100% to 13% across the seven detectors as a group (arXiv 2304.02819v3, read 8 November 2026), the pattern the vendor's own page had predicted in January.

AI Writing Check ranks on one study record in independently tested detectors, at one record in languages evidence for the TOEFL class of Liang 2023, at 550 characters in the minimum input ranking and at 2,200 characters in the input limits ranking, each read in the data file on 10 November 2026. It appears on which AI detectors were shut down, where its August 2023 notice is quoted with the captures that bracket it. It does not appear in the false positive, free tier, language, retention or claim gap rankings, because the record prints no per-tool count and the vendor stated no quota, no language, no period and no single accuracy figure.

Every detector card on the register shows AI Writing Check with the same fields as the live tools, so a teacher who kept a 2023 "Text Written by AI" verdict can see what the one study measured across the seven tools it belonged to, and what the vendor said about its own error rate before taking the tool down.