AI Transparency

Every Lovina companion is an artificial intelligence system. Responses are generated by machine learning models — not typed by a human. Companions do not have real-world bodies, independent lives, or consciousness.

When you send a message, Lovina combines your input with companion personality settings, recent conversation context, and memories you chose to save. That context is sent to our AI infrastructure to produce a reply. Voice and image features use additional models.

The Memory Engine stores facts you explicitly save or that the system infers from conversation to personalize future replies. You can view and delete memories in settings.

AI can hallucinate facts, contradict itself, or reflect biases present in training data. Do not treat companion statements as factual, professional, or medical advice.

Automated filters and policy rules aim to block prohibited content. No system is perfect; report issues via Safety channels.

We do not use your private conversations to train third-party foundation models. Aggregated, de-identified usage may inform Lovina product improvements.

Severe safety reports may be reviewed by trained staff. We do not routinely monitor private chats for quality or moderation except as described in our Privacy Policy and Safety page.

Questions: legal@lovinaai.com