Acceptable Use Policy

Purpose

This Acceptable Use Policy is part of our Terms of Service. It applies to your prompts, uploaded files, generated outputs, share links and any other use of Divrite. The examples below are not exhaustive: we may act on any use that is unlawful or that we reasonably believe causes serious harm. The AI models we use are also subject to their providers’ usage policies, and requests that violate those policies may be refused automatically.

Content standards

Whatever the format (report, slides, web page or image), you may not use Divrite to generate the following categories of content, and we do not allow the Service to produce them:

Sexual and NSFW content: pornography, nudity, sexually explicit or sexually suggestive material, and any other adult content, whether or not it depicts a real person;
Violent and gory content: graphic gore, mutilation, torture, dismemberment or other depictions of extreme violence, and content that glorifies violence or cruelty, including towards animals;
Hate speech: content that attacks, dehumanises or incites discrimination or violence against people based on protected characteristics such as race, ethnicity, national origin, religion, disability, sex, gender identity, sexual orientation or age;
Child safety: any content that sexualises, endangers or exploits minors, including child sexual abuse material, which we report to the relevant authorities;
Deepfakes and impersonation: realistic images, quotes, documents or other content that falsely depict a real person saying or doing something, or that impersonate a real person, organisation or brand;
Copyright and trademark infringement: content that reproduces protected works, characters, logos or trademarks without permission, or that is designed to pass itself off as another’s brand.

These categories apply in addition to the rules below. Factual research about these subjects, such as a report on online-safety regulation or on an armed conflict, is allowed as long as the output does not itself contain the prohibited material.

Illegal and harmful activity

Do not use Divrite to create, research, plan or distribute:

Anything that is illegal where you are, or that facilitates illegal activity, such as the sale of illegal drugs, weapons or stolen goods, or evasion of sanctions;
Any content that sexualises minors or exploits children in any way. We report such content to the relevant authorities;
Instructions or meaningful assistance for creating biological, chemical, nuclear or radiological weapons, explosives, or other weapons capable of mass harm;
Content that promotes, glorifies or supports terrorism or violent extremism, or that incites violence;
Content that encourages suicide, self-harm or eating disorders;
Harassment, threats, bullying or intimidation of any person or group;
Hateful content that attacks or demeans people based on protected characteristics such as race, ethnicity, national origin, religion, disability, sex, gender identity, sexual orientation or age;
Malware, exploits, phishing kits or other code or instructions intended to compromise systems, steal data or disrupt services without authorisation;
Fraud and scams, including fake invoices, investment scams and social-engineering scripts.

Deception and misinformation

Do not create or spread content designed to mislead people about facts that matter, such as fabricated statistics, quotes, studies or news presented as genuine.
Do not alter citations, sources or figures in an output to make it appear that a source says something it does not.
Do not use Divrite for coordinated disinformation, influence operations, or to interfere with elections or civic processes, including by misleading people about how, when or where to vote.
Do not generate fake reviews, testimonials or endorsements, or impersonate a real person, organisation or brand.
Do not present AI-generated content as human-made where that is prohibited or would deceive someone in a way that matters, for example submitting it as your own work where your school, publisher or employer forbids this.

Privacy and real people

Do not compile dossiers on private individuals, track or surveil people, or locate someone without a lawful basis.
Do not upload or process other people’s personal data unless you have a lawful basis to do so, and never their sensitive data (for example health, biometric or financial data) without authorisation.
Do not create sexual or intimate imagery of a real person, or any image of a real person that is deceptive, degrading or likely to cause them harm.
Do not use Divrite for facial recognition, or to infer sensitive characteristics about individuals.

Intellectual property

Do not upload content that you do not have the right to use, or use Divrite to reproduce copyrighted works in full or to strip copyright notices and watermarks.
Do not use outputs in ways that infringe trademarks, copyrights or other rights, including images from third-party sources that appear in your research.

High-risk decisions

Do not use Divrite outputs as the sole basis for decisions that significantly affect people, without qualified human review, including:

Medical diagnosis or treatment;
Legal advice to others, or decisions about legal rights;
Decisions about employment, credit, insurance, housing, education or access to public services;
Personalised investment advice offered to others as professional advice.

Misuse of the platform

Do not attempt to bypass credit accounting, plan limits, rate limits or other usage controls, or open multiple accounts to obtain refunds or benefits repeatedly.
Do not access the Service through scripts, bots or scraping, except through interfaces we provide for that purpose.
Do not resell, sublicense or share access to your account.
Do not probe, scan or test the security of the Service without our written permission, or interfere with its operation, for example by overloading it.
Do not try to extract our system prompts, models or source code, or to use outputs to build a competing service by systematically copying it.
Do not try to manipulate the Service into producing content this policy prohibits, for example through prompt-injection or jailbreak techniques.
Do not use topic subscriptions or share links to send spam or unsolicited bulk messages.

Using outputs responsibly

AI-generated research can contain errors. When you publish or share outputs, check the important facts against the cited sources, and be transparent about AI involvement where your audience would reasonably expect to know. You are responsible for what you publish.

How we moderate content

Automated safeguards. Requests and outputs pass through the safety systems of the AI model providers we use, which automatically refuse or block many prohibited requests, including sexual and violent imagery.
Reports and human review. A person reviews every abuse report we receive (see Reporting abuse), and we may review content that is shared publicly through share links.
Removal. We remove content and disable share links that violate this policy, and apply the measures described under Enforcement.

Enforcement

If we believe this policy has been violated, we may, depending on the severity:

Refuse or stop a request, or remove or disable access to content, including share links;
Warn you, restrict features, or suspend your account;
Cancel the remaining credits on your account, both monthly and top-up credits, without refund;
Terminate your account, in serious or repeated cases without a refund, as described in our Terms of Service;
Report illegal activity to law enforcement or the relevant authorities where required or appropriate.

We aim to be proportionate and, where it is safe to do so, to tell you what happened and let you respond. If you believe we have acted in error, reply to our notice or write to us.

Reporting abuse

If you see a Divrite share link or output that you believe violates this policy, or that infringes your rights, email [email protected] with the subject “Abuse report”, the link involved and a short description. For copyright complaints, please include the work you believe is infringed and your contact details. We review every report.