Working safely with AI

Usage & Privacy Notice

Explained transparently: processing, safeguards and your responsibility when using MEDIAN AI.

Back to the portal

MEDIAN GROUP | AI is the MEDIAN Group's in-house AI access. All requests go through a central gateway operated by us — not through public consumer services. This notice explains transparently where and how your inputs are processed.

1. Where your data is processed

The central entry point is a gateway operated by the MEDIAN Group, running on our own hardware in a German data center. Authentication, access control, usage metering and rate limiting take place exclusively there.

Our own models

Some of the language models run entirely on our own hardware. For these, inputs never leave the data center.

Claude (Anthropic)

Claude models are provided via Amazon Web Services (Amazon Bedrock) — exclusively within the EU (see below).

2. Claude via Amazon Bedrock — EU-resident

For Claude models, only our gateway calls the Amazon Bedrock cloud service. The connection is technically hard-limited to data centers within the EU; requests to locations outside the EU are rejected. Processing takes place at the following EU locations:

3. No training use, no prompt logging at the cloud provider

4. Access & authentication

5. Logging & telemetry

For billing, capacity management, quality assurance and troubleshooting, usage and telemetry data is collected and stored on our own infrastructure. Depending on the purpose, this may also include the content of requests and responses. This data never leaves our infrastructure and is not shared with third parties.

6. Your responsibility as a user

This notice describes the processing at a glance. In addition, the group policies and the data-protection requirements of the MEDIAN Group apply. If you have questions about the processing of personal data, please contact the data protection officer.