Mistral may use your Vibe conversations and uploads for AI training unless you opt out
3 sept. 2026 · Collecté automatiquement à partir de sources publiques. Non vérifié par une personne.
Ce contenu n'est disponible que dans sa langue d'origine (anglais).
What happened
Mistral AI's help center states that input and output data in its products — conversations, uploaded documents, and other user-provided content — "may be included in Mistral's model training programs." For the consumer Vibe app (formerly Le Chat), users are not opted out by default: training continues until the user disables the toggle labeled "Allow your interactions to be used to train our models" (Admin panel → Manage → Vibe → Privacy), or, on mobile, deselects "Enable data sharing" under Settings → Data & Account Controls. Documents attached or uploaded within Vibe count as input data. Vibe Enterprise customers are opted out of training by default, and the Vibe and Studio/API toggles are separate — each must be configured on its own. Feedback (thumbs up/down and accompanying comments) may be used for training regardless of the other settings.
Why this matters for creators
Anything pasted or uploaded into a consumer Vibe session — drafts, scripts, lyrics, manuscripts, client files — can become model-training data unless the toggle is off. Mistral's documentation notes that the opt-out applies only going forward; it does not remove data already used.
How to opt out
- Web/desktop: Admin panel → Manage → Vibe → Privacy → disable "Allow your interactions to be used to train our models".
- iOS/Android: Settings → Data & Account Controls → deselect "Enable data sharing".
- Studio/API: separate toggle in the Admin panel under Privacy ("Anonymous improvement data").
Evidence note
The help article carries no visible publication date. Its current opt-out wording was publicly documented in early September 2026 (a 2026-09-03 research digest noted the help-center update) and re-verified against the live page on 2026-09-28.
Sources
- Can I opt out of my input or output data being used for training? — Mistral AI (primary)
- Mistral AI's Training Data Defaults Spark Privacy Pushback Among Enterprises — WebProNews (secondary)