Safety & wellbeing
Your safety matters to us. This page provides important information about using LUKi responsibly and resources if you need support.
LUKi is an AI companion and may not be suitable for some minors.
Understanding LUKi
LUKi is an AI, not a human. While LUKi is designed to be a friendly companion, it's essential to understand that all responses are generated by artificial intelligence. LUKi does not have feelings, consciousness, or the ability to truly understand your situation.
What LUKi is:
- ✓An AI-powered conversational companion for casual chat and entertainment
- ✓A creative tool for brainstorming, writing, and exploring ideas
- ✓A fun way to pass time and engage with AI technology
What LUKi is not:
- ✗A replacement for human relationships or professional support
- ✗A medical, legal, financial, or mental health professional
- ✗A source of factual information that should be relied upon without verification
- ✗An emergency service or crisis intervention tool
AI limitations & accuracy
LUKi uses advanced AI technology, but like all AI systems, it has significant limitations:
- 1May generate incorrect information: LUKi can produce responses that sound confident but are factually wrong. Always verify important information from authoritative sources.
- 2Cannot access real-time information: LUKi doesn't browse the internet or know current events unless specifically provided.
- 3May produce harmful or biased content: Despite safety measures, AI can occasionally generate inappropriate, biased, or harmful responses.
- 4Cannot truly understand context: LUKi processes text patterns but doesn't genuinely understand emotions, nuance, or your personal situation.
Healthy usage guidelines
- •Balance AI interaction with real relationships. LUKi should complement, not replace, human connection.
- •Take breaks. If you find yourself spending excessive time with LUKi, consider stepping away and engaging with other activities.
- •Don't share sensitive personal information. Avoid sharing passwords, financial details, or highly personal information.
- •Seek professional help when needed. If you're struggling with mental health, relationships, or other serious issues, please consult qualified professionals.
- •Remember it's entertainment. Treat LUKi as a fun AI tool, not as a real friend or therapist.
Information for parents & guardians
A person under 18 can use LUKi only with the approval of a parent or legal guardian. You approve from your own account and stay in control in Family. If your child uses LUKi:
- •Have open conversations about AI and its limitations
- •Monitor usage and ensure it doesn't interfere with real-world activities
- •Discuss the difference between AI companions and real relationships
- •Ensure your child understands not to share personal information
How LUKi Checks Age
We ask every person for their birth month and year, once. Afterwards it can be changed in the app only to a later birth month, which means a younger age, when the device's age settings rule out the one given. We do not check identity documents.
A person under 18 cannot create an account alone. They send a request to a parent or legal guardian, who approves it from their own account with a paid plan. For a child under 13 we also confirm a completed payment.
On iPhone and Android, where the law requires it, LUKi also asks the device for the age range its owner or a parent set. If that range cannot be true with the birth month given, we hold the account until a later birth month is given, a parent approves, or the person shows us that the device is wrong. The device tells us a range, never a date of birth, and we share it with nobody.
If you believe a child is using LUKi without a parent's approval, write to us. We will lock or delete the account.
How LUKi Responds to Crisis Language
An automated check reads each message as it is sent and looks for signs that the person may be in danger, such as talk of suicide or self-harm. It has three parts. The first is a list of phrases in the 11 languages LUKi supports, built on the categories of the Columbia suicide severity screener. The second is the AI model that writes the Character's reply. As it answers, it also judges whether the message speaks of the person's own suicidal thoughts or self-harm. The third is a safety model run by our AI provider, which reads the same message.
When the check finds a risk, LUKi adds crisis helpline details to the reply, in the language in use. LUKi adds them itself, so they are there whatever the Character writes. In the apps and on the website each helpline also appears as a button that calls it in one tap. Every Character is instructed to respond with care and never to encourage self-harm, and a reply that the safety model reads as self-harm content is replaced.
The check is automatic. No person reviews it. For an adult it alerts nobody, and we keep no score, label or record from it against the account. We keep one number: how many times a day helpline details were shown, with nothing that identifies a person. For a child's account, the parent who approved it gets an email saying that a crisis helpline was shown. The email does not say what the child wrote. From 1 July 2027 we also keep a copy of that child's conversation for 3 years, as California law requires. Nobody else is alerted, including the authorities.
The check cannot be switched off, and it applies to every Character and every user.
LUKi is not a crisis service. The check will miss real emergencies, and it will sometimes react to words that were not meant seriously. If you or someone else is at risk, contact your local emergency number or a crisis line directly.
Report issues & get support
If you encounter harmful, inappropriate, or concerning AI responses, please report them to help us improve:
If you're in crisis
If you or someone you know is in immediate danger or experiencing a mental health crisis, please contact emergency services or a crisis helpline immediately. LUKi is not equipped to handle emergencies.
By using LUKi, you acknowledge these safety guidelines and agree to our Terms of Service.