Ceci est un projet antérieur au lancement. Il peut évoluer après relecture juridique.
Ce document n’est pas encore disponible dans votre langue ; l’original anglais est donc affiché. La version coréenne ou anglaise prévaut.
AI Interpretation Notice
Dernière mise à jour :
Lingolane's interpreted speech and captions are generated by AI. They can be wrong, so double-check what matters and tell the people in your meeting that you use AI interpretation.
1. We use generative AI
Lingolane interprets speech with OpenAI's real-time speech translation model. The interpreted speech the other person hears, the interpreted speech you hear and the captions on screen are all AI-generated.
Interpreted speech uses a synthetic AI voice. It does not imitate or clone your voice.
2. Limits of accuracy
- Meaning can change or be dropped. Pay special attention to numbers, dates, amounts, names and technical terms.
- When several people talk at once or audio is quiet, interpretation may lag or skip.
- Do not rely on interpretation alone for contracts, legal, medical or safety decisions; confirm them in writing or with a human interpreter.
3. Tell the people in your meeting
We recommend telling the other party before the meeting starts. The interpreter screen lets you copy a notice in 13 languages to paste into the meeting chat.
Example: This meeting uses AI real-time interpretation. It may not be perfectly accurate, so please confirm important points in the chat or in writing.
4. How audio and records are handled
- For interpretation, meeting audio streams from your browser to OpenAI (USA) in real time. Lingolane servers do not store audio.
- OpenAI states that it does not use data received through its API to train models by default.
- Caption history (the transcript) is saved only on your computer.
See the Privacy Policy for details.
5. Contact and reports
If interpretation caused a problem or you notice inappropriate output, please tell us at [email protected].