Use of AI Policy

Last updated: November 2025

This Use of AI Policy describes how Hello Ada employs artificial intelligence technologies within:

  • helloada.ai
  • edu.helloada.ai

The policy explains how AI models are integrated, what data they process, and the controls we have implemented to ensure safety, privacy, and compliance, particularly in school and municipal environments.

1Purpose of AI in Hello Ada

Hello Ada uses AI to support learning, creativity, and prototyping. Examples include:

  • Generating code prototypes from student ideas
  • Providing guided learning and explanations
  • Offering feedback, refinement suggestions, and structured prompts
  • Assisting teachers with classroom workflows

AI is used as a supportive tool, not as a replacement for teaching or human judgement.

2AI Models and Providers

Hello Ada uses the following AI model providers:

  • Anthropic (Claude)
  • Google Cloud (Gemini)
  • SkoleGPT (Nordic education-focused LLM)

All AI providers are configured in no-training and no-retention modes, meaning:

  • Data is not used to train or improve their models
  • Data is not stored beyond what is required to return a response
  • Providers process information solely to generate outputs requested by the user

3What Data Is Sent to AI Providers

Hello Ada may send the following information to AI providers:

  • The prompt written by the user
  • Relevant short chat history required to generate a coherent response

edu.helloada.ai (educational platform)

  • Students are instructed not to input personal data
  • The platform does not require any identifying information
  • Use is supervised by teachers
  • The school is responsible for ensuring students do not enter personal details

Both Platforms

  • Hello Ada does not intentionally send personal data to external AI providers
  • If a user chooses to include personal information in a prompt, it may be processed temporarily by the AI model to generate an output
  • Hello Ada includes safeguards to discourage and limit such behaviour

4Safeguards and Safety Controls

To support safe and compliant use of AI, Hello Ada employs the following controls:

Technical Controls

  • No-training, no-retention AI configurations
  • EU-region processing where available
  • Schema-restricted tool calls
  • Basic automated redaction filters
  • Input validation to discourage personal data entry
  • Output safety checks

Organisational Controls

  • Students use the platform under teacher supervision
  • Clear guidance to students not to include personal information
  • Access to AI functionality by children is limited to non-identifying, project-based content
  • A Data Processing Agreement governs the EDU environment

5How AI Outputs Should Be Used

AI-generated outputs are suggestions, prototypes, or drafts. They may contain:

  • Inaccurate information
  • Gaps in logic
  • Code requiring correction or refinement

Educators remain responsible for guiding students and evaluating AI-generated content. AI outputs should not be considered authoritative, complete, or final.

6Prohibited Uses

Users (including students) may not use Hello Ada's AI features for:

  • Entering or processing identifiable personal data
  • Generating harmful, unsafe, or discriminatory content
  • Circumventing school policies or digital safety requirements
  • Creating content not suitable for children
  • Any purpose that violates applicable law or school rules

7Transparency for Students and Educators

We aim to promote responsible and transparent use of AI in classrooms. Participants should understand:

  • When AI is being used
  • How AI assists in learning
  • That AI outputs are generated automatically
  • That teachers remain the primary point of guidance and authority

8Data Storage and Retention

Data provided to AI models is not stored by model providers beyond the time it takes to generate a reply.

Within Hello Ada:

  • Student project data is retained for 90 days
  • Logs are retained 30–90 days
  • Backups are retained for 30 days

These strict retention periods ensure compliance with GDPR principles of data minimisation and storage limitation.

9Updates to AI Systems

AI capabilities and safeguards may evolve as technology develops.

Any material changes to the AI tools or their data processing characteristics will be reflected in this policy with an updated "Last updated" date.