Privacy & safety
AI Companion Privacy and Safety: A Practical Guide
Simple habits for safer companion conversations: share less sensitive data, verify important claims, and keep human boundaries clear.
Explore AmaviCheck the product before you start sharing
Companion chats can feel informal and personal, which makes it easy to disclose more than you intended. Before starting, review the service privacy policy, account controls, and available ways to delete, reset, or report content.
Look for plain explanations of what data is collected, why it is used, how long it may be kept, and which service providers may process it. A friendly interface is not a substitute for those disclosures.
Treat generated replies as unverified
AI systems can confidently produce false, outdated, or invented information. Emotional tone does not make a factual claim more reliable. Verify important information with authoritative sources and qualified people.
A companion should not be the sole basis for medical, legal, financial, emergency, or personal safety decisions. If you or someone else may be in immediate danger, contact local emergency services or an appropriate crisis resource.
Keep emotional and time boundaries visible
An AI companion is designed to respond, but it does not have human needs or independent feelings. You do not owe it continued conversation, secrecy, money, or obedience. You can pause, reset, or leave whenever you choose.
If use begins to interfere with sleep, responsibilities, spending, or relationships you value, set a time limit and speak with someone you trust. The goal is an experience you control, not one that controls your routine.
Use account, memory, and reporting controls
Secure the email or identity provider connected to your Amavi account and sign out of shared devices. When a service uses password login, choose a unique password. Review the available memory settings and reduce or disable conversation context when you do not want earlier exchanges to shape future replies.
Report content that appears unsafe, illegal, exploitative, or inconsistent with platform rules. A report gives the service a chance to investigate the system behaviour; it does not require you to continue the conversation.
Common questions
Frequently asked questions
Are AI companion conversations completely private?
Do not assume any online service is private in an absolute sense. Review the applicable privacy policy and controls to understand processing, storage, service providers, and legal disclosures before sharing sensitive information.
Can I trust advice from an AI companion?
Treat it as generated, unverified content. Check important claims against authoritative sources and use qualified human help for medical, legal, financial, emergency, or safety decisions.
What should I do if a reply makes me uncomfortable?
Stop the interaction, retry or reset if you want to continue, and use the product reporting tools when the content may violate platform rules. You never need to keep engaging with the character.