Data privacy
We prioritize data safety and user privacy. In this section, we outline how data is managed and the measures taken to protect user privacy when using the AI Assistant.
To enable the AI Assistant capabilities, table and column names from the current database schema are transmitted to AI provider. This step is crucial for accurately translating user requests into SQL queries. Key considerations regarding data privacy are as follows:
- Data sharing options: By default, only table and column names are sent to AI provider.
-
User Consent: On the first use of AI completion for a specific connection, DBeaver will prompt for your confirmation to send metadata. This confirmation is mandatory to use the feature.
-
Log Transparency: The entire request can be logged for your review. To enable this, navigate to Preferences and check the Write GPT queries to debug log option.
- Selective Metadata Sharing: If you prefer not to share information about certain tables, adjust the tables in scope using the Scope field.
Privacy policy¶
In addition to these general principles, each AI provider applies its own privacy policy:
| Provider | Privacy policy |
|---|---|
| OpenAI | Requests are processed under the OpenAI privacy policy. |
| GitHub Copilot | Copilot follows the GitHub privacy statement. |
| Azure OpenAI | Operates under its own privacy policy. It's recommended to review their terms. |
| Google Gemini | Data is handled according to Gemini privacy measures. |
| Ollama | Ollama follows a local-first design: all AI processing happens on your machine or a self-hosted/company server, and prompts, code, or documents never leave your computer. |
| Anthropic Claude | Requests are processed according to Anthropic’s privacy policy. Review their terms to understand how your data is handled. |