Subprocessors
The complete list of subprocessors is included below.
Anthropic
We use the Anthropic API for the Claude 3 Sonnet chat model.
This model is optional within the AI-School environment.
Anthropic has a DPA that fully complies with GDPR regulations.
Google Cloud
We use Google Cloud for database, cloud storage, and application backend services (Firebase).
We use Google Cloud Vertex AI for the Gemini chat models.
By using these services, we are subject to the Google Cloud data processing agreement.
Google is committed to ensuring its services comply with GDPR, and the Cloud Data Processing Addendum is part of this effort.
We use servers within the European Economic Area. Details can be found in the appendices of the Data Processing Agreement.
Google for Education
We also offer login via Google for Education.
We only request basic profile information to process the login.
Microsoft
We use Microsoft Entra to log in employees and students with their school account.
This login method is secure because no passwords are involved. After successful login with Microsoft, we receive the necessary information to complete the login process.
We only request rights to share basic profile information with us to log in the user.
Mistral
We use the Mistral API to call the Mistral Large chat model.
This model is optional within the AI-School environment.
Mistral has a DPA that fully complies with GDPR regulations.
We do not send personal data to the Mistral API.
Moneybird
We internally use the Moneybird accounting package to process invoices.
Within Moneybird, we fill in personal data of contact persons within the school.
We do not process data of other employees or students within Moneybird.
OpenAI
We use the OpenAI API to call the chat models GPT 3.5 Turbo and GPT 4.0 and the image model Dall E 3.
We also use the Assistant API to set up Assistants (custom chatbots).
We have signed a data processing agreement with OpenAI that complies with GDPR regulations.
OpenAI does not process personal data from AI-School. We do not send personal data to the OpenAI API.
Replicate
Replicate platform provides API connections for calling the open source model Llama 2.
We fall under Replicate's privacy terms for calling APIs.
We have assessed these terms as compliant with GDPR.
These open source models are optional within the application.
We do not send personal data to Replicate.
Serper
We use Serper.dev to perform internet searches within our chat functionality. Serper collects and logs certain technical information about API requests, including IP addresses, timestamps, and searches. This data is used to monitor platform performance, security, and functionality.
Serper offers the ability to anonymize searches by masking the 'query parameter' in their logs. They do not process personal data of students or employees, and the collected data is only used for technical purposes such as debugging and performance improvement.
Stability AI
Stability AI provides API connections for calling the open source models SDXL and SD3.
We fall under Stability AI's privacy terms for calling APIs.
We have assessed these terms as compliant with GDPR.
These open source models are optional within the application.
We do not send personal data to Stability AI.