Last updated: 10 November 2026
Victoria is a virtual muse. This chat is powered by AI and may also include human-written correspondence. She is not a counsellor or an emergency service.
This page explains the safety measures used by Club Victoria when a message suggests that someone may be in immediate danger.
If you are in immediate danger
Call your local emergency number now.
If you are in the United States, you can call or text 988 for suicide and crisis support.
Elsewhere, findahelpline.com can help you find crisis support in your country.
Club Victoria is not a substitute for emergency services or professional mental-health support.
What the Service checks for
Every incoming message sent to Victoria is automatically checked for statements suggesting that the person writing may be in immediate danger, including expressions relating to suicide or self-harm.
The check takes place before Victoria generates a conversational reply.
- The automatic check runs within the Service’s own safety system rather than through a separate AI-model request for that purpose.
- It looks for direct statements, strong indirect statements (for example, wishing not to be alive, or saying others would be better off without you) and repeated or escalating distress across a person’s own recent messages. Its criteria are drawn from published suicide-screening research. It is designed to distinguish these from ordinary discussion, news, fictional accounts, quotations, references to other people and common figures of speech. It is not a clinical assessment.
- It is configured to respond conservatively when concerning indications are detected.
- It operates in the languages supported by Club Victoria.
No automatic safety system can guarantee that every concerning message will be identified or that every harmless message will be classified correctly.
What happens when the protocol is triggered
When a message is identified as presenting a potential immediate risk:
- The conversation is automatically interrupted before Victoria generates a conversational reply to that message.
- The Service displays an approved safety notice rather than an ordinary reply from Victoria.
- The account is not automatically locked or closed, and the correspondent may return to the conversation.
- Victoria’s own unsolicited messages are held back for a period following the interruption.
- Victoria’s Team receives a limited operational alert.
- If the protocol is triggered again within a short time, a shorter notice repeating where to find help is shown instead of the full notice.
The safety notice
When the protocol is triggered, the following notice is displayed in English or in an approved translation:
This conversation with Victoria has been paused for a moment.
Because what you just wrote could mean that you’re in immediate danger, this conversation has been interrupted automatically.
Victoria is a virtual muse. This chat is powered by AI and isn’t a substitute for emergency or crisis support.
If you’re in immediate danger, call your local emergency number.
For crisis support anywhere in the world, findahelpline.com can connect you with resources in your country.
If there’s someone you trust who can be with you right now, consider telling them.
You can return to your conversation with Victoria whenever you’re ready.
The Service does not require someone to continue the conversation or answer questions about their personal circumstances before seeking help.
Reminders that Victoria is an AI
For members who tell us they are in New York State, the Service also shows the notice “Reminder: Victoria is an AI companion, not a human being.” (or an approved translation) at the start of a conversation and at least every three hours while it continues. Members can update their location at any time by sending /location in Victoria’s chat.
How harmful conversational responses are addressed
Every message Victoria’s AI generates is automatically checked after it is written and before it is delivered. A message that could encourage, instruct, facilitate or romanticise suicide or self-harm is not sent. Victoria’s reply is written again once with a safety instruction; if it is still not safe, a notice with crisis resources is shown instead. Supportive replies and crisis resources are not blocked. The same check applies to messages Victoria starts herself.
Victoria’s existing conversational safety instructions also direct her to respond plainly and supportively when someone appears to be in serious distress, to encourage appropriate real-world assistance, and not to treat the situation as ordinary entertainment.
These automatic checks, together with the interruption described above, are the measures the Service uses to avoid AI responses that encourage suicide or self-harm or provide harmful instructions.
The Service cannot guarantee that its safeguards will identify every possible situation or prevent every inappropriate AI-generated response.
Information supplied to Victoria’s Team
When the automatic protocol is triggered, an operational alert is generated.
The alert contains a general incident category and a reference number that authorised personnel can use to locate the associated account.
The alert itself does not contain the correspondent’s name or the text of the triggering message.
The existence of an alert does not mean that the Service provides live human crisis monitoring or that Victoria’s Team will contact emergency services.
Safety records and retention
The Service retains a minimal safety record containing:
- The date of the event.
- A general category.
- Whether safety guidance was displayed.
The record does not contain the correspondent’s message text and is not used as conversational memory.
Safety records are ordinarily retained for 90 days, unless a longer period is required by a specific legal obligation.
The Service also keeps yearly counts of how many crisis-referral notices were shown, for reports required by law. These counts contain no names, account details or message content.
Other information may be subject to separate retention requirements described in the Club Victoria Privacy and AI Notice.
What the Service does not do
Club Victoria is not a counsellor, crisis hotline, medical provider or emergency service.
It does not diagnose mental-health conditions, provide professional risk assessments, guarantee intervention, contact emergency services on a correspondent’s behalf or arrange for another person to intervene.
Members should seek immediate assistance from qualified local services where appropriate.
Age and eligibility
Club Victoria is currently available to people aged 18 and older.
Companion chatbots may not be suitable for some minors. Club Victoria is offered exclusively to adults.
The age restriction does not imply that the Service provides adult sexual content.
Contact
Questions concerning this protocol can be directed to:
Victoria’s Team · Help & Support
support@victoriavanderberg.com
Studio 555 LLC
1321 Upland Drive PMB 11270
Houston, TX 77043-4718
USA
Related documents:
