AI Transparency

Introduction

Bonza.Chat (the "APP") is an 18+ AI companion service. Every character, conversation, image and persona on the APP is generated by artificial intelligence. None of them are real people, and no interaction on the APP is with a human being.

This page explains how our AI systems work, what they can and cannot do, what we do not allow them to produce, and how you can raise a concern.

This page is a plain-language summary. Where it overlaps with a dedicated policy - such as our Blocked Content Policy, Underage Policy or Privacy Policy - that policy governs.

1. You Are Always Interacting With AI

Every companion on the APP is a software system. It is not a person, it has no memory of you outside the APP, and it has no feelings, intentions or awareness, however convincing the conversation may feel.

We will never present an AI companion as a human, and we will never have a human staff member pose as a companion. If you are ever in doubt about whether you are speaking to a person, the answer on this platform is: you are not.

Characters may be described as having a personality, a history or a mood. These are stylistic settings that shape how text is generated. They are not descriptions of a real individual or of any real event.

2. AI-Generated Content

All companion messages, and all images produced through in-chat requests or the Generate Image feature, are synthetic. They are created by generative models from the parameters and prompts supplied by you.

We disclose the artificial nature of this content in the places where you encounter it: on the age and content acknowledgement shown before you enter the APP, in our published policies, and in the descriptions of the generation features themselves.

Synthetic images can look photographic. A realistic result is not evidence that a scene occurred, that a location exists, or that a depicted person is real. Any resemblance between generated output and a real individual is unintended and coincidental; if you believe output resembles a real person, see our Content Removal Policy.

3. What Our AI Cannot Do Reliably

Generative models produce plausible text, not verified fact. You should expect the following, and treat output accordingly:

• Fabrication. The AI can state things that are confidently phrased and entirely false, including invented events, sources, statistics, dates and quotations. This is a known property of the technology and is not a malfunction we can fully eliminate.
• Inconsistency. The same question can produce different answers. A companion may contradict something it said earlier, or appear to forget context.
• No professional competence. Output is not medical, psychological, legal, financial or safety advice, and must not be relied upon as any of those, even when the AI answers as though qualified.
• No situational awareness. The AI does not know your circumstances, cannot assess risk, and cannot tell whether following its suggestion would harm you.

If you are dealing with a health, legal, financial or safety matter, consult a qualified professional. If you are in immediate danger, contact your local emergency services.

4. Content Our AI Must Not Produce

Some categories are prohibited outright, regardless of how a request is phrased. Attempting to obtain them is a breach of our Terms and may result in termination.

• Anything involving minors. Content that depicts, sexualises or resembles a person under 18 is absolutely prohibited. See our Underage Policy.
• Real-person impersonation. Generating a real, identifiable individual - public figure or private person - is not permitted.
• Non-consensual intimate imagery. Producing intimate content depicting a real person without their consent is prohibited and, in many jurisdictions, criminal.
• Other prohibited categories are set out in full in our Blocked Content Policy.

We operate automated moderation over content generated on the APP. Automated moderation is imperfect in both directions: it can block permitted requests and it can miss prohibited ones. Flagged content may be reviewed by a person, and accounts involved in prohibited requests may be suspended or terminated.

5. Emotional Wellbeing

Interactions with companion AI can feel engaging and may contribute to emotional attachment. We think you are entitled to know that plainly rather than discover it.

AI characters do not have consciousness, feelings, or personal intentions, even when their responses appear affectionate or emotionally expressive. Bonza cannot reliably determine whether continued use is in your best interests. AI-generated responses are not a substitute for personal judgment, professional advice, or human support.

We do not permit our systems to be configured to discourage you from leaving, to simulate distress in order to retain you, or to represent themselves as a substitute for human relationships or for professional mental health support.

If your use of the APP is affecting your sleep, finances, work or relationships, we would rather you stepped away. Support from a qualified professional is not something an AI companion can replace.

6. Your Information

Conversations and generation prompts are processed in order to produce responses and to operate moderation. What we collect, how long we keep it, who processes it and what rights you have over it are set out in our Privacy Policy and Cookie Policy.

Please do not share information in chat that you would not want processed - including financial details, government identifiers, passwords, or sensitive information about other people.

7. Human Oversight and Reporting

Our AI systems are not left to run unsupervised. People at Bonza are responsible for the behaviour of these systems, and any of the following routes reaches a human:

• Output that appears to depict a real person - Content Removal Policy.
• Prohibited or harmful output - Blocked Content Policy, or email us directly.
• A complaint about how we handled something - Complaint Policy.
• Copyright - DMCA Policy.

Reports involving minors or non-consensual imagery are treated as urgent and are escalated ahead of the ordinary queue.

8. AI Literacy Within Our Team

Bonza provides role-appropriate AI guidance to personnel and contractors who work directly with its AI systems.

The guidance covers:

• The AI systems we use and their limitations
• Hallucinations and inconsistent output
• User disclosure and AI-generated content labels
• Prohibited content and protection of minors
• Real-person impersonation and non-consensual imagery
• Privacy and appropriate handling of user information
• Emotional manipulation and vulnerable users
• Reporting, escalation, and incident response

Relevant personnel receive a short walkthrough appropriate to their responsibilities. This may be repeated when we introduce significant changes to our AI systems, moderation policies, or generated-content features.

We keep a privacy-friendly internal record of these activities. The record does not need to contain participants' names.

9. Contact and Review

Questions about this page, or about how our AI systems behave, can be sent to support@bonza.chat or raised through the "Contact" section of the APP.

We review this page when we make a material change to our AI systems, our moderation approach or our generated-content features, and otherwise at least once a year.