Votre vie privée
Français

Lieux à contacter

Mistral AI

Type
AI company
Lieu
France — France / EU
Dernière vérification
2026-09-22
Prochaine vérification
2027-03-21

Pas encore traduit — affiché en anglais.

Comment les joindre

  • Content / usage policy report form · Web form
    Contact à accès limité · 2026-09-22

    Qui peut l’utiliser: users-meeting-published-eligibility

    Restrictions: 21 categories of content violation (hate speech, CSAM, violence, NCII, fraud, IP). Intro: 'Mistral AI is committed to maintaining a safe environment for all. This form allows you to report content you believe violates our terms or applicable laws.'

    Preuve que ce moyen accepte les prises de contact · 2026-09-22

    21 categories of content violation (hate speech, CSAM, violence, NCII, fraud, IP). Intro: 'Mistral AI is committed to maintaining a safe environment for all. This form allows you to report content you believe violates our terms or applicable laws.'

    Content moderation, not model safety. Requires you to supply a URL or a shared private chat link to the offending content — so it cannot be used to report a general model flaw.

  • HackerOne vulnerability submission · Responsible disclosure programme
    Contact à accès limité · 2026-09-22

    Qui peut l’utiliser: users-meeting-published-eligibility

    Restrictions: Security vulnerabilities. Published on the contact page, which notes 'General bugs shall be reported via our standard support channels'.

    Preuve que ce moyen accepte les prises de contact · 2026-09-22

    Security vulnerabilities. Published on the contact page, which notes 'General bugs shall be reported via our standard support channels'.

    No stated AI/model safety scope.

  • Press · Email address
    Contact à accès limité · 2026-09-22

    Qui peut l’utiliser: users-meeting-published-eligibility

    Restrictions: Media enquiries only.

    Preuve que ce moyen accepte les prises de contact · 2026-09-22

    Media enquiries only.

    Listed here only because it is one of the few published addresses.

  • Homepage · Homepage
    Vérification expirée

    Preuve que ce moyen accepte les prises de contact

  • No published frontier safety framework or responsible scaling policy found · Published policy or framework
    Vérification expirée

    Preuve que ce moyen accepte les prises de contact

Ce qu'il fait

No publicly named safety or alignment function. Trust Center covers data protection and security compliance; the Usage Policy is enforced through a content-reporting form.

Évaluation franche

No public accounts found of anyone reaching a Mistral safety function. No safety@ or disclosure@ address is published anywhere on mistral.ai or legal.mistral.ai — checked the contact page, terms, and usage policy directly. The report form's requirement of a content URL structurally prevents reporting a capability or alignment finding.

Notes

For a European frontier developer this is a thin public surface: not found — general contact only, at https://mistral.ai/contact. If you have a model-safety finding, the HackerOne form is the only route with a human triage attached, but it is framed as security. No safe-harbour language and no whistleblower channel found.

Sources

Une erreur ?