Our Anti-Humanization Policy

What the Anti-Humanization Policy Bans

Tools must be named because vagueness allows bans to be worked around. We have undetectable.ai, StealthGPT, WriteHuman, HIX Bypass, Phrasly, and BypassGPT which will all be prohibited, along with QuillBot in any form of their rewriting, and Spinbot spinners. So will the bypass or stealth mode in a product that isn’t otherwise concerning.

It also bans the practice regardless of brand. New humanizers are released every month, so the clause is written around behavior. It bans any process that has the purpose of changing text so that detection software gets a different reading. This captures manual synonym switching to change the score, as well as running a draft through three rewriters in order.

The surprising thing is that the ban on human writing is enforced against every author. An author who has produced absolutely every word of a text may avoid running the text through a humanizer if a detector has flagged it. This transforms a false positive into an automatically determined case where no one can defend the text, and it destroys the record of the drafts that is the author’s only defense.

• Undetectable.ai, StealthGPT, WriteHuman, HIX Bypass, Phrasly, BypassGPT.
• QuillBot rewriting modes, Spinbot, and any article spinner.
• Bypass or stealth settings inside otherwise ordinary tools.
• Rewriters chained together until a detector score comes down.
• Manual synonym-swapping carried out purely to move a number.
• Humanizing a genuinely human draft that a tool flagged by mistake.

Why the Anti-Humanization Policy Matters

Using a draft that contains open AI is a disclosure problem. Hiding that it was AI is a deception problem and problems of this nature should not overlap. A buyer should have the option to decide how to respond if informed that a draft was AI assisted. A buyer finds out later that text submitted to a buyer was laundered text. This buyer has no choice but to make false claims about the text.

There is craft damage on top of the ethics. Text is humanized by avoiding repetition of predictable words. This often involves using less precise terms, broken idioms and other minor imperfections which appear to balance the text. It also includes the insertion of non-semantic words that sound more natural. The worst case is regulated copy. A clause that describes what a product may do is altered to describe what the product actually does. This conveniently shifts the burden to consumers for identification until someone does.

Laundering diminishes evidence quality for everyone. After fraudulent practices become the norm, an honest supplier’s product report holds little value. Then an honest writer’s clean record holds little value, and a buyer must rely on a brand name. Given the situation of evidence loss of value, we prefer the evidence retains its value more, therefore this policy is stricter than the AI ban it is part of.

How the Anti-Humanization Policy Protects Human Writing

Our writers gain firsthand experience. With plain, clear writing, especially in technical and instructional writing formats, the algorithms are more likely to find evidence of academic dishonesty. Those who are doing the most honest work are the ones who are more likely to be flagged. A policy secretly allowing humanizers would encourage those who write in a clear way to process their drafts through plagiarism checkers.

Evidence is used to evaluate flags instead of flags themselves. When a file that we believe to be human has returned a reading that comes back high, the editor retains the wording in the draft, captures notes and previous drafts, and the transcription log and sends the editor the report alongside their explanation. No one is asked to modify the writing to appease any utilities or score. No one is asked to edit their writing to self-impose a high score.

The effect for you is that a report from us means exactly what it says. You are not looking at a fabricated number, and when a number is not convenient, you still see it. That is significantly smaller than the claim that all documents score clean, and that is a claim we can realistically keep.

• Flagged human drafts are defended with evidence, never rewritten to order.
• Awkward detector readings are sent to you instead of being disregarded.
• There’s never a request for writers to revise anything with the intention of achieving a higher score.
• Draft history is retained so a false positive can be answered.
• Confirmed humanizer use ends the engagement the same day.

Anti-Humanization Policy FAQs