No, our AI models are not trained on your data. This is documented in our privacy policy under “Transparency in AI use”: “Prompts or other text generated through the platform are not used to train or improve the models.”

This means that when you use our tools to translate, summarize, or proofread copyrighted material, the text is:

• Only used to deliver the immediate service
• Not part of the model’s training data
• Not available to other users
• Automatically deleted after processing

You can therefore safely use all our AI assistants with copyrighted material without worrying about copyright infringement.

All your data stays within the EU:

• Web hosting: Denmark (Simply.com)
• AI processing: Sweden (Microsoft Azure OpenAI data center)
• No data transfers to countries outside the EU

Yes, SpellBright is fully GDPR compliant. We:

• Are subject to Danish and EU data protection law
• Have a dedicated Data Protection Officer (DPO)
• Only use GDPR-certified partners (Simply.com and Microsoft Azure OpenAI)
• Have carried out a data protection impact assessment
• Implement privacy-by-design principles
• Comply with all GDPR requirements on data minimization, pseudonymization, and user rights

We are recommended for Danish educational institutions, which underlines our focus on safe use in the education sector.

• SpellBright (SkrivSikkert ApS): Has NO access to your prompts or AI-generated outputs
• Microsoft Azure OpenAI: Processes data only to deliver the service – no access in normal operation
• Other users: Have NO access to your data

We have comprehensive procedures in place for handling data breaches:

• Immediate assessment of the scope of the breach
• Notification of Datatilsynet within 72 hours
• Direct communication with affected users
• Measures to limit the consequences
• Thorough investigation and implementation of preventive measures

However: Because of our data minimization practice (no permanent storage of prompts/outputs), the risk of a meaningful data breach is extremely low, as we simply do not hold data that can be lost.

We have extensive security measures in place:

Technical measures:

• Encryption of data during transmission and storage
• Anonymization through encrypted identifiers
• Multi-factor authentication for access to sensitive systems
• Regular security updates
• Automatic content filters and abuse monitoring

Organizational measures:

• Strict access controls
• Ongoing staff training in data protection
• Regular security assessments and audits
• Dedicated DPO (Data Protection Officer)

We have designed our platform specifically with the education sector and the protection of minors’ data in mind:

No personal identification:

• No user logins on the main platform
• No collection of personally identifiable information
• Full anonymity for all users

Local data processing:

• All user activity is processed only in the browser’s local memory
• Data does not leave the user’s device (apart from encrypted communication with the AI model)
• Automatic deletion at the end of the session

Special protection for minors:

• No profiling or tracking of user behavior
• No risk that pupils’ or students’ work becomes visible to others
• Teachers and students can use the tool without worrying about GDPR issues

This ensures that the platform can be used safely in all educational settings, whether primary school, upper secondary school, or higher education.