Will editing software make my writing look like AI?
Wretify is a clean, minimalist writing app for nonfiction writers, built for focus — with an AI that has read everything you've written and refuses to write it for you.
The short answer
Correcting spelling and grammar in your own sentences does not turn them into AI writing: the words, choices and structure are still yours. Letting an AI rewrite whole sentences or generate new ones is different — the text is then partly the model’s, and that is what AI detectors are built to look for. Grammarly itself separates the two: its grammar and spelling AI “does not change the substance of the writing”, while its generative AI “can generate content and meaningfully change writing”.
The other half of the answer is that AI detectors are not reliable enough to settle the question on their own, for anyone.
How accurate are AI detectors?
Less accurate than most people assume. OpenAI, the maker of ChatGPT, withdrew its own AI classifier on July 20, 2023, “due to its low rate of accuracy”: it caught 26% of AI-written text and wrongly labelled human writing as AI 9% of the time.
A Stanford study (Liang and colleagues, 2023) ran seven detectors on 91 essays written by non-native English speakers for the TOEFL exam. On average the detectors labelled 61% of them as AI-generated; at least one detector flagged 89 of the 91. The same detectors were almost perfectly accurate on essays by US eighth-graders.
Turnitin, whose detector is widely used in education, reports a document-level false positive rate below 1% for documents with 20% or more AI writing, and about 4% for single sentences — and says it “cannot mitigate the risk of false positives completely”, and that detectors are “resources, not deciders”. Grammarly says that no detector today “can reliably and conclusively determine whether AI was used to produce text”.
How do I keep my writing clearly my own?
- Use checkers to correct, not to rewrite: accept spelling and grammar fixes, and decline suggestions that rephrase whole sentences.
- Turn off generative features you do not want. Grammarly, for example, lets you switch its generative AI off.
- Keep your drafts and their history. Versions of a document written over days are the strongest evidence of authorship there is.
- If you are a non-native English writer, know that the research above found detectors biased against you, and keep your drafts all the more.
What does Wretify do?
Wretify’s AI is built to refuse to write for you. It thinks with you, asks questions, and brings you your own earlier passages and real research, but it does not draft your chapters, and nothing reaches the page unless you press Replace or Insert, or drag it there yourself. Its grammar pass makes the smallest change that fixes what is broken, leaves your words and style alone, and tells you what was wrong so you learn it. Your chapters are plain files in your own Google Drive.
Sources
- OpenAI — New AI classifier for indicating AI-written text (updated July 20, 2023) — openai.com/index/new-ai-classifier-for-indicating-ai-written-text/
- Liang et al. (2023), GPT detectors are biased against non-native English writers — arxiv.org/abs/2304.02819
- Turnitin — Understanding the false positive rate (2023) — www.turnitin.com/blog/understanding-the-false-positive-rate-for-sentences-of-our-ai-writing-detection-capability
- Turnitin — Debunking common myths (2024) — www.turnitin.com/blog/does-turnitin-detect-ai-writing-debunking-common-myths-and-misconceptions
- Grammarly — How can I turn off the Grammarly AI? — support.grammarly.com/hc/en-us/articles/26993719119501-How-can-I-turn-off-the-Grammarly-AI-Does-Grammarly-use-trigger-AI-detectors
Updated . Prices and features of other apps are taken from their own websites on that date and can change. Wretify is our product, so we have tried hard to be fair to the others. If something here is wrong or out of date, tell us and we will correct it.