Data privacy and AI

A common question: is my data used to train AI models? The answer is no, never.

No training on your data

  • Conversations: your exchanges with AI models (Assistant, Explore, etc.) are never used to train or improve models.
  • Documents: files you upload to Drive are never transmitted to AI model providers for training.
  • Translations: texts processed by Trad are neither stored nor reused by the models.

How Delos protects your data

Isolated architecture

Each organization has an isolated environment. Data from one organization is never accessible to another.

No AI-side persistence

When you send a request to an AI model via Delos:

  1. Your request is sent encrypted to the provider
  2. The model generates a response
  3. The response is returned to you
  4. The provider retains no data from this interaction

We have negotiated specific contracts with all model providers (OpenAI, Anthropic, Mistral, etc.) guaranteeing non-retention and non-use of your data.

Retention and deletion

  • Conversations: kept as long as your account is active. You can delete them individually at any time.
  • Documents: kept until manually deleted by the user.
  • After account deletion: all data is purged within 30 days.

Model transparency

For each model available on Delos, we indicate:

  • The provider (OpenAI, Anthropic, Mistral, etc.)
  • The processing location
  • The contractual confidentiality guarantees