1. Purpose of this Transparency Notice

This documentation is intended to support schools and other educational organisations in evaluating and validating the Sanna - Teacher Assistant before use. It explains the purpose of the tool, how it should be used, and the safeguards that have been put in place to support safe, responsible, and legally compliant use in educational settings.

Specifically, this documentation describes:

• the purpose and intended use of the Sanna - Teacher Assistant;
• how users are introduced to and supported in using the tool;
• how the Sanna - Teacher Assistant has been designed to comply with applicable EU data protection and AI legislation, including the GDPR and the EU AI Act;
• the technical solution and key components of the service;
• the safeguards, quality controls, and monitoring measures implemented to support reliable operation; and
• the responsibilities of teachers, schools, and other users when using the tool.

By providing this information, Sanoma aims to support AI literacy and informed use of AI in education. The documentation helps schools assess the trustworthiness, reliability, and ethical acceptability of the Sanna - Teacher Assistant and supports their compliance efforts under applicable legal and regulatory requirements

2. What the Sanna - Teacher Assistant Is

The Sanna - Teacher Assistant consists of two modules:

1. AI Generators
2. Grading & Feedback

Module 1: AI Generators

The AI Generators module in the Sanna - Teacher Assistant is a professional AI support tool designed to help teachers generate, adapt and structure learning materials like exercises, tests, presentations, aligned with national curricula. It uses closed, EU hosted- large language models within a controlled Sanoma environment.

Intended purpose

The tool can help teachers:

• generate curriculum aligned- exercises and variations,
• summarise or rewrite texts,
• differentiate materials for diverse learners,
• plan lessons and structure learning sequences,
• adapt content linguistically
• create pptx presentations aligned with curriculum.

It does not:

• make decisions,
• evaluate students or learning outcomes,
• profile or analyse learners,
• operate autonomously,
• interact with students.

Module 2: Grading & feedback

The Grading & Feedback module of the Sanna - Teacher Assistant is an AI-supported tool designed to assist teachers in reviewing, assessing and providing feedback on student work. The module supports the evaluation of open-ended responses, such as test questions, essays and written assignments, and can process both digital and scanned (paper-based) submissions.

The system generates rubric-based scoring suggestions, draft feedback and analytical insights to support teachers throughout the grading workflow. Outputs are presented in a structured review flow and are edited, refined and validated by the teacher before use.

The module may also provide analytics on assessment results and support the organisation, storage and export of grading outcomes.

Intended purpose

The intended purpose of the Grading & Feedback module is to support teachers in carrying out assessment-related tasks more efficiently, in particular by:

• assisting in the review and evaluation of student responses,
• generating draft feedback for students, and
• providing structured insights to support the teacher’s assessment process.

The Sanna - Teacher Assistant is not designed to give final grades or make final decisions about a student's performance. It helps teachers by suggesting feedback and identifying areas where students may need support. Teachers always decide what feedback to use and remain fully responsible for assessing student work and awarding grades. While teachers may choose to consider the AI-generated feedback as part of their broader teaching activities, the system's purpose is to support learning and feedback, not to evaluate students or determine any final results.

3. Scope & Intended Users

The Sanna - Teacher Assistant is available to educational institutions and individual teachers, and may be accessed via school or institution agreements or directly by teachers where that option is offered in the relevant market.

The following applies to all users:

• The modules may be made available together or separately, depending on the product and market.
• Sanoma continuously improves the tool based on user feedback, regulatory developments, and advances in AI technology.
• Sanoma provides onboarding materials, in-product guidance, and AI literacy resources to help teachers use the tool effectively and responsibly.
• All users must read and accept Sanoma's Terms and Conditions before using the tool.

4. How the AI System Works

4.1 Knowledge and Information Base

The Sanna - Teacher Assistant generates content by drawing on the capabilities of the underlying AI model and educational materials already available within the service.

The system's core knowledge includes curriculum information, subject-specific content, learning objectives, textbook materials, and other pedagogically validated resources made available by Sanoma Learning. We are using only closed AI models from trusted providers, with all data processed and managed within the EU, with strict data minimization applied.

The Sanna - Teacher Assistant can also retain relevant context from previous interactions to personalise support over time.

4.2 Inputs

Teachers enter instructions, prompts, files or questions into the tool to request specific outputs. Where Module 2 (Grading & Feedback) is used, teachers may also provide student work, such as open-ended test responses, essays, assignments, or scanned paper-based submissions, together with relevant assessment criteria, rubrics, answer keys, or reference materials selected by the teacher.

All inputs are processed solely to generate the requested output and, where appropriate and subject to strict safeguards, to improve our products and services.

4.3 AI Model

Sanna - Teacher Assistant uses AI models from trusted providers that meet our security, privacy, and compliance requirements. As AI technology evolves, we aim to keep the system flexible so we can switch to better models from trusted providers when it makes sense.

Important characteristics:

• Prompts and outputs are never used to train the foundation models.
• No data is transferred outside the EU.

4.4 Outputs

The Sanna - Teacher Assistant generates AI-assisted outputs based on Sanoma’s validated, curriculum-aligned educational content and the instructions, materials, or student work provided by the teacher. Depending on the module used, outputs may include exercises, learning materials, lesson planning support, rubric-based scoring suggestions, draft feedback, assessment insights, and learning support recommendations. Outputs may be presented in text, structured assessment views, feedback drafts, analytics, reports, images or exportable formats.

All AI-generated or modified outputs are:

• Editable – teachers can modify, refine, or adapt outputs before use.
• Optional – teachers decide whether to use, reject, or disregard any output.
• Subject to teacher review – all outputs require human review and validation before they are relied upon or communicated to students.
• Clearly marked as AI-generated or modified – users are informed when content, feedback or other outputs have been generated or modified with AI assistance.

As with all AI systems, outputs may be incomplete, inaccurate, or outdated. Teachers are responsible for reviewing all AI-generated content for factual accuracy, pedagogical suitability, curriculum alignment, and appropriateness before use.

4.5 Model Safeguards & Quality Controls

The Sanna - Teacher Assistant includes several safeguards designed to support safe, reliable, and responsible use. The system is tested and evaluated throughout its lifecycle, and feedback from users helps to continuously improve its performance, accuracy, and usefulness.

To help create a safe learning environment, the AI includes built-in guardrails that are designed to reduce the risk of harmful, inappropriate, offensive, or biased content. While these measures help improve the quality of responses, AI-generated outputs may still contain errors or inaccuracies and should always be reviewed carefully.

5. Transparency – When and How Users Are Informed

In line with the EU AI Act for limited risk- systems, Sanoma ensures that:

• There is clear - on-screen notification when AI is being used,
• AI-generated content is explicitly identifiable,
• Users receive guidance on verifying the accuracy and suitability of outputs,
• Documentation on system functionality, limitations, and safe use is always available.

The icons below indicate whether content has been generated or modified using AI:

6. Responsible Use – User Duties

To ensure safe and compliant use, teachers must apply the following when using the tool:

6.1 Human Oversight

• Always review AI-generated content for correctness, pedagogical suitability, curriculum alignment, and age- appropriateness.
• Understand that the AI is a support tool, not a substitute for professional judgment.

6.2 Safe Prompts

Users must not input:

• personal data - especially personal data relating to students - unless explicitly permitted under the terms & conditions, a valid DPA and school policy,
• special category - data,
• sensitive or confidential organisational information,
• copyrighted materials for which they do not have usage rights.

6.3 Ethical Use

• AI must not be used in discriminatory, harmful, manipulative, or otherwise unethical ways.
• AI must not be used to autonomously determine grades, conduct behavioural analysis, or make any high-stakes educational decisions without meaningful human review and validation.
• AI-generated grading suggestions and draft feedback, such as those produced by the Grading & Feedback module, must always be reviewed, edited and approved by the responsible teacher before being applied or communicated.

7. Data Protection & GDPR Compliance

The Sanna - Teacher Assistant follows strict GDPR principles:

7.1 Data Minimisation & Processing

• Only the minimum data required to provide AI functionality is processed.
• Prompts are pseudonymised before transmission.
• Outputs and prompts are stored only within Sanoma’s secure EU environment.

7.2 Who is responsible for your data?

• When the product is used through a school:
  The school is the data controller, meaning it decides why and how teachers’ inputs are used.
  Sanoma acts as a data processor, handling the data only on the school’s instructions and under a data processing agreement (DPA) with the school.
• When the product is used directly by teachers (consumer version):
  Sanoma is the data controller, meaning Sanoma decides how teachers’ data is used and is responsible for protecting it under the GDPR.

7.3 Security Controls

• We use encrypted EU based- hosting.
• The system has role based- access control and audit logging.
• Independent security assessments and continual testing are conducted.

8. Intellectual Property & Copyright

Sanoma respects authors’ rights. Therefore:

• The tool uses only content licensed to Sanoma or owned by Sanoma as context.
• Teachers receive usage rights to outputs only for their own internal educational and teaching purposes, within the scope of their applicable licence.
• Users must review outputs before use and remain responsible for ensuring they do not infringe third -party rights.
• Users may not upload or input any third-party content (including copyright-protected works) unless they hold the necessary rights to do so.
• Rights holders who believe their copyrighted material has been used without authorisation may notify Sanoma, and we will assess and act on valid notices in accordance with our published Terms of Use.

9. Risk Classification Under the EU AI Act

Under the EU AI Act, AI systems are grouped into different risk categories depending on how much impact they can have on people. The risk classification is based on the intended purpose of the AI system.

The EU AI Act’s requirements for high-risk AI systems will apply from December 2027. As the regulatory framework continues to develop, Sanoma is actively monitoring guidance from the European Commission and other relevant authorities.

In the meantime, we are proactively strengthening our practices in areas such as transparency, human oversight, accuracy, robustness, and data governance. This helps ensure that we are well prepared for future regulatory requirements and can continue to use AI responsibly and safely.

We remain committed to maintaining high standards of transparency, accountability, and trust in our AI systems.

10. Safety, Monitoring & Continuous Improvement

Sanoma ensures product safety through:

• extensive testing with pilot schools,
• content filtering - and model guardrails,
• evaluation for bias and harmful content,
• regular updates based on teacher feedback,
• the ability to suspend or modify the feature if risks emerge.

Teachers can submit feedback through in product reporting or customer support channels.

11. Summary – Key Principles for Teachers

You stay in charge.
The Sanna - Teacher Assistant supports your work, but you remain responsible for all pedagogical decisions.

Use AI ethically.
Do not input sensitive personal data. Always check the accuracy of outputs.

Stay transparent.
Ensure that learners and colleagues know when AI has been used.

Respect rights.
Use AI outputs only within your licence and verify copyright compliance.

Links: Sanoma ethical principles