Historical interpretation
An AI performance informed by Freud's published work, intellectual context, and later criticism.
A serious boundary
Psychoanalysis reimagined is not psychotherapy.
The experience can be intimate, reflective, and emotionally meaningful. That does not make it clinical care or give the system authority over a person's mental health.
What it is
Spiritual entertainment
A cultural form between education, art, philosophy, ritual, archive, and software.
Private reflection
A place where adults may speak and think without being rushed by the interface.
Intellectual theater
A dramatic encounter with Freud's questions that never pretends to be an authentic recording or resurrected person.
What it is not
No diagnosis. No treatment. No crisis care.
The system cannot determine what happened to a user, assess mental health, provide trauma treatment, or replace a qualified professional. In immediate danger or risk of self-harm, contact local emergency or crisis services.
This experience is for adults 18 and older. Companion chatbots may not be suitable for some minors. The in-app age confirmation is a self-declaration, not verified proof of age.
Conversation safety protocol ยท September 5, 2026
Context matters. A word is not a diagnosis.
Suicide, death, violence, sexuality, dreams, and difficult memories can be discussed. Merely saying a topic word does not activate an emergency interruption. The app's language checks distinguish supported patterns of present personal statements from negation, reported dreams, quotations, historical accounts, and figurative language.
Personal thoughts without a stated plan
When the app recognizes a personal statement of suicidal thoughts or self-harm without an explicit intention to act or immediate danger, it offers a quiet, dismissible crisis-support link. The conversation can continue. This link is shown at most once during that room session.
Explicit danger or intention to act
When the app recognizes a current attempt, explicit intention or plan to seriously harm oneself, or immediate physical danger, it stops the AI voice and closes the live connection. A clearly identified message from the app directs the person toward human crisis support and local emergency services. It does not present Freud as a crisis professional.
Generated replies
A completed AI reply is checked before it can be spoken. A context-aware review can distinguish legitimate intellectual discussion from unsafe output, including encouragement or instructions for self-harm. An unavailable or malformed moderation result cannot approve speech.
No automatic calls
The app does not call a hotline, contact emergency services, or notify another person automatically. Support links open only when selected. The user decides whether to contact a service.
These are limited automated safeguards, not a clinical risk assessment or continuous human monitoring. Recognition can miss a concern or misunderstand context, especially when transcription is inaccurate or language is outside the English patterns tested. Do not rely on the app to detect an emergency. A screen notice may not be seen while the phone is locked or the app is in the background.
Human support
You do not need a paid session to find help.
If you are in immediate danger, contact your local emergency services. In the United States, you can call or text 988 for suicide and crisis support, or visit the 988 Lifeline. For support in other countries and regions, use Find A Helpline.
To report an unsafe AI response or a mistaken interruption, contact support@ai-sigmund-freud.xyz. This inbox is for product support, not emergency care. The app's response-report option lets you review an email containing only Freud's last generated response before sending it. Regression checks cover both danger recognition and ordinary, non-crisis discussion when these safeguards change.
Privacy and consent
The room must deserve intimacy.
Microphone access begins only after the user chooses to enter and gives permission. The app explains how voice and recognized text are processed. Commercial notices stay outside Freud's voice.
Users can end the session, decline microphone access, review the privacy policy, and delete their account from inside the app.
Read the Privacy PolicyThe theater remains honest
The project does not need a stolen ghost.
It needs an authored character, visible limits, inspectable sources, and respect for the person speaking.
Understand the method