Posting guidelines

These guidelines exist so that someone arriving from a search engine finds a clear question, a clear answer and enough evidence to judge it. They are short on purpose. If something is not covered here, ask yourself whether the post helps the next reader.

Asking a question

Put the real question in the title. "Is a 38% AI score evidence of cheating?" is better than "Help with detector". In the body, say what you already know, what tool or situation you are dealing with, and what decision you need to make. Remove names, student numbers and anything else that identifies a private person.

Search first. Many questions about false positives, watermarks and deepfake tells have already been answered, and a reply to an existing thread keeps the evidence in one place.

Writing an answer

Answer the question in the first two sentences, then explain. Readers skim, and answer engines quote the opening of an answer, so the opening should stand on its own.

Back up claims about accuracy, bias or performance with a source. Link the study, standard or institutional statement itself. If you are sharing your own experience, say that it is your experience and what it covered, for example "in my course of 120 essays last term". Personal experience is welcome. It just should not be dressed up as a measurement.

Be precise about uncertainty. "This detector produced false positives on non-native writers in a 2023 study" is useful. "Detectors never work" and "detectors are always right" are not.

Links

Links to sources, papers, standards and documentation are welcome. Links from new accounts are shown as plain text until the account has a few approved posts, which keeps spam off the page without stopping you from citing anything. Links added by members are marked so search engines do not treat them as endorsements.

Do not post links whose main purpose is to sell a product or service. Mentioning a tool you used is fine. Dropping a referral link in every answer is not.

What gets removed

• Attempts to identify, accuse or shame a private person.
• Content that helps someone evade detection in order to deceive a teacher, employer or publication.
• Non-consensual synthetic imagery of real people, in any form.
• Harassment, slurs and personal attacks.
• Spam, including repeated self-promotion.

How moderation works

Posts can be reported by any member. Reports go to a moderation queue that staff review. When a post is removed, its author sees the reason and can appeal. Moderation actions are logged, and staff do not delete posts simply because they disagree with them.

New accounts go through a short trust period during which their posts are reviewed before everyone can see links in them. It usually ends after a handful of approved posts.

Staff posts

Forum staff wrote the first threads here so that the forum is useful from day one. Every staff post carries a Staff label. Staff answers follow the same sourcing rules as everyone else and can be challenged the same way.

Corrections

If an answer is wrong or out of date, reply with the correction and your source. If it is your own answer, edit it and say what changed. A visible correction is more useful to the next reader than a silent one.