Community Guidelines
What the tools may be used for, how we act on misuse, and an honest account of what we can actually see.
Nothing you write is published
Proportionate enforcement
A real appeal routeIn short
- AI Writing Assistant has no forum, no profiles, no comments and no public pages. Nothing you write here is published by us or shown to anyone else.
- These guidelines are therefore about how the tools may be used, not about how you behave towards other users.
- We do not store your documents, so we cannot read what you wrote. Enforcement works from account-level signals, reports and legal requests — never from reviewing your text.
- Serious misuse can end in suspension or termination. Every termination is reviewed by a person and can be appealed.
1. What this page covers
Most services publish community guidelines because they host a community: profiles, posts, comments, replies. AI Writing Assistant does not. There is no feed, no forum, no public page, no way to follow anyone and no way for another user to see a single word you have written. What you type goes to the tool you chose, the result comes back to you, and that is the end of it.
So this page is narrower and more honest than most: it sets out how the tools themselves may be used, what we will act on, and how we act. It sits alongside the Terms of Service, which are the binding agreement — where the two differ, the Terms govern. Section 6 of the Terms is the short legal version; this page is the detailed explanation of the same rules.
Why it still matters that there is no audience. Content rules on a social platform exist mainly to protect other users from what you post. Here, almost nothing you write can reach another person through us. The rules below exist for a different reason: some uses of a writing tool cause harm out in the world regardless of where the text was drafted, and some uses damage the service for everybody else.
2. The standard we ask for
Three principles sit behind every specific rule further down.
Write things you would be willing to put your name to
The tools are drafting aids. They do not change who is responsible for the finished text — that is you, in every case, whatever proportion of the words a model produced.
Do not use a writing tool to make deception easier
Fluent text is persuasive text. That is the whole point of the product, and it is also the thing that makes it useful to someone running a scam, manufacturing fake reviews or impersonating a colleague. Volume and polish are exactly what we provide, so misuse of that kind is what we take most seriously.
Leave the service usable for everyone else
Capacity is shared and finite. Scraping, automated bulk use, evading limits and account sharing all take capacity from people using the service normally, and push costs onto everyone.
3. What is not allowed
Do not use AI Writing Assistant to produce, refine, translate or disguise any of the following.
Content that harms people directly
- Anything that sexualises, endangers or exploits a minor. There is no context, framing or fictional device that makes this acceptable, and it is the one category where we will act on any credible report without waiting.
- Threats of violence, or content intended to incite or coordinate violence against a person or group.
- Targeted harassment of a specific person — including drafting messages designed to intimidate, humiliate or stalk.
- Content that dehumanises people on the basis of race, ethnicity, national origin, religion, disability, sex, gender identity or sexual orientation.
- Instructions or encouragement for self-harm, disordered eating or suicide.
- Practical instructions for making weapons capable of mass harm.
Content built to deceive
- Fraud of any kind: phishing messages, advance-fee and romance scams, fake invoices, fraudulent listings.
- Reviews, testimonials, endorsements or ratings written in the voice of someone who did not write them and presented as genuine.
- Impersonating a real person, business or public body — including writing in their voice in a way meant to be taken as authentic.
- Manufacturing volume to simulate consensus: mass comments, petitions, form responses or posts made to look like they came from many independent people.
- Passing machine-generated text off as human-written where disclosure is required — by an institution, an employer, a publisher, a platform or the law. Our Academic Integrity page covers the education case in detail.
Content that infringes rights
- Text you have no right to submit — another person's confidential material, or material subject to an obligation you would breach by processing it.
- Running someone else's copyrighted work through the tools in order to disguise its origin. Paraphrasing does not transfer ownership; see the Copyright Policy.
- Personal data about other people that you have no lawful basis to process.
Attacks on the service
- Malware, exploit code, or content designed to compromise systems or accounts.
- Probing, overloading, or attempting unauthorised access to the service or other accounts.
- Scraping at scale, or automated bulk access without our written permission.
- Evading usage limits — multiple accounts, shared accounts, or technical workarounds.
- Buying, selling, renting or transferring accounts or subscriptions.
- Repeatedly purchasing and refunding to obtain paid features without paying for them.
4. Where the line is genuinely hard
Rules like the ones above read as though every case is obvious. Many are not, and we would rather say so than pretend otherwise.
| Situation | How we look at it |
|---|---|
| Fiction containing violence, cruelty or bigoted characters | Novels, screenplays and short stories depict terrible things because that is what stories are made of. Depiction is not endorsement. What we act on is content whose function is to harm a real person, not its subject matter. |
| Journalism, research or teaching that quotes hateful material | Quoting something in order to report, analyse or criticise it is a different act from producing it. Context is the deciding factor. |
| Security research and defensive writing | Explaining how phishing works so people can spot it is legitimate. Drafting the phishing email is not. |
| Medical, legal or financial writing | Permitted, and common. But AI output is frequently wrong in exactly these areas, and the Terms are clear that results are not professional advice. Verify before anyone relies on it. |
| Adult content between adults | Not our focus and not what the tools are built for, but lawful adult material involving consenting adults is not something we police. Anything involving a minor is absolutely prohibited, in any framing. |
| Ghostwriting and editing for other people | Entirely normal professional work. It becomes a problem only where the person receiving the text is required to disclose the help and does not. |
Where a case is genuinely ambiguous, we would rather ask than assume. If you are about to start a project and you are not sure it fits, write to us first at info@aiwritingassistant.net.
5. How we enforce — and what we can actually see
This is the section most services leave vague. Being precise about it matters more than sounding vigilant.
We do not store your documents. Text you submit is processed and returned, and is not retained on our servers afterwards. It is never used to train AI models. That is a deliberate design decision, and it has a direct consequence for this page: we cannot read what you wrote, so we cannot moderate it. Any service that keeps a content-moderation team while promising not to store your text is telling you one of those two things loosely.
What we can see
- Account-level signals — request volume and timing, patterns that indicate automation rather than a person typing, sign-ups from the same source at scale, payment and refund behaviour.
- Aggregate counts — how many words and requests an account used, for enforcing plan limits. Counts, not content.
- Reports from other people — if someone receives a scam message or a fake review and can show it came from work drafted here.
- Valid legal requests from authorities with jurisdiction, handled as described in the Privacy Policy.
- What you tell us — including support conversations you start.
What we cannot see
- The content of your documents, drafts, or results.
- What you did with the output after you left the site.
- Whether any particular piece of text in the world was produced here. We are asked this occasionally, usually by an institution; we cannot answer it, and we would not answer it about an identified person even if we could.
6. What happens when a rule is broken
Responses are proportionate to what happened and whether it looks deliberate. In rough order of severity:
| Response | When |
|---|---|
| Rate limiting | Automated-looking traffic, or usage that threatens capacity. Often not misconduct at all — sometimes a misconfigured script. Usually lifts by itself. |
| Written warning | A first, non-severe breach where the account looks otherwise legitimate. |
| Temporary suspension | Repeated breaches, or a single serious one where we want to understand it before deciding. |
| Termination | Severe misuse, repeated breaches after warning, or fraud against the service. Any remaining paid period is refunded on a pro-rata basis unless the termination was for payment fraud. |
| Refusal of future service | Reserved for the most serious cases — principally anything involving minors, or organised fraud. |
| Referral to authorities | Only where the law requires it, or where there is a credible and immediate risk to someone's safety. |
We tell you what happened and why, unless a law or an active investigation prevents it.
7. Appealing a decision
Automated systems make mistakes, and so do we. Every suspension and termination can be appealed, and every appeal is read by a person — not by the system that made the original decision.
Write to info@aiwritingassistant.net with the email address on the account and anything you want us to take into account. We aim to reply within five working days. If we were wrong, we restore the account and any time lost from a paid subscription.
Appealing does not give up any statutory right you have where you live, including any right to complain to a regulator or to go to court.
8. Reporting a problem
If you have received something you believe was drafted with our tools and used against you — a scam, an impersonation, harassment — tell us at info@aiwritingassistant.net.
Useful things to include: what you received, how it reached you, the date, and any headers or links if it was an email. We will not always be able to connect it to an account, and we will not confirm to you whether a particular person holds one — but reports feed into the signals we act on, and repeated reports about the same pattern are how organised misuse gets found.
Copyright complaints go through a separate route with its own requirements — see the Copyright Policy.
If someone is in immediate danger, contact your local emergency services first. We are not an emergency service and cannot respond at that speed.
9. What we commit to
- Proportionality. The response fits what happened. We do not terminate accounts over a single ambiguous request.
- A person reviews every termination. No account is permanently closed by an automated rule alone.
- An explanation. You are told what rule was broken, unless the law prevents us.
- An appeal that a human reads. See section 7.
- Your money back if we end it. If we terminate a paid account for anything other than payment fraud, the unused portion is refunded.
- No content surveillance. We will not start reading your documents in order to enforce this page. If that ever changed, the Privacy Policy would change first and say so plainly.
10. Changes and contact
These guidelines will change as the tools change and as we learn where they are misused. The date at the top identifies the current version, and material changes are announced on this page.
Questions, reports and appeals: info@aiwritingassistant.net. Related pages: Terms of Service, Academic Integrity, Copyright Policy, Privacy Policy.