## AI at Coursebox

center #faf7f2 AI Policy AI built to support learning, with people always in control.

Coursebox uses AI to help course creators work faster, learners stay engaged, and training teams do more with less. We believe AI should make good learning more accessible, not replace the human expertise that makes learning meaningful.
## AI as your assistant, not your author

left #F4EDF7 Human in the loop Every AI-generated piece of content in Coursebox — whether a course outline, a quiz question, or a video script — is created as a starting point for a human to review, edit, and approve. Nothing reaches a learner until a course admin or creator has decided it should.

That is not a feature we might add later. It is how Coursebox works by design.
## Our commitments

center #faf7f2 Commitments
#faf7f2 #F7F0FB Course content, learner data, quiz responses, and any material you create or upload is never used to train or fine-tune AI models. Not by Coursebox, and not by any of our AI providers. This is contractually assured across every model we use, including Microsoft Azure OpenAI, Google Gemini Pro, and HeyGen. Your content never trains our AI From August 2026, Coursebox will include AI disclosure notes on interactive AI features like the AI Tutor, machine-readable markers on AI-generated content, and contextual disclosures on AI-generated video. Learners deserve to know when they are interacting with AI. We are transparent about AI with learners Every AI provider Coursebox integrates is evaluated for security, data handling, and responsible AI standards before we work with them. We use Microsoft Azure OpenAI rather than the public OpenAI API specifically because it provides stronger data protection and no risk of your content being used for model training. We choose our AI providers with care AI can draft a course. It cannot decide whether that course is accurate, appropriate, or right for your learners. That decision always belongs to the person who publishes it. Humans stay accountable
## Limits we have set on purpose

center #F4EDF7 Safeguards
#F4EDF7 #faf7f2 AI will not generate sexually explicit, violent, or otherwise inappropriate content. No inappropriate content AI features must not be used to create misleading, defamatory, or discriminatory content. No misleading or discriminatory content The AI Tutor is designed with safety parameters that, by default, keep it within the scope of the course it is assigned to. It will not give learners answers to quiz questions or assessments. Course creators can choose to open the AI Tutor to answer broader questions outside the course material, but the default behaviour is intentionally restricted to protect the integrity of the learning experience. AI Tutor is course-scoped by default Coursebox does not use AI to make automated pass or fail decisions about learner performance without human review. No automated pass/fail decisions
## Preparing for the EU AI Act

center #faf7f2 EU AI Act From 2 August 2026, Coursebox will comply with EU AI Act Article 50 obligations. Article 50 requires providers of AI systems that interact with people, and systems that generate AI content, to be transparent about when AI is involved.
#faf7f2 #F7F0FB Learners interacting with the AI Tutor will be informed they are engaging with an AI system, in line with Article 50 requirements for AI systems that interact with natural persons. AI Tutor disclosures AI-generated course content, including text, images, and video, will carry machine-readable markers identifying it as AI-generated, as required under Article 50. AI content markers AI-generated avatar videos and AI voiceovers will include contextual disclosures so learners understand the content they are watching was produced with AI tools. For courses delivered via SCORM, LTI, or iframe outside the Coursebox platform, the course publisher is responsible for ensuring appropriate learner-facing disclosures are in place. AI video disclosures
## Questions or concerns?

center #F4EDF7 Security & Privacy false externalReference false link If you believe an AI feature in Coursebox has produced inaccurate, biased, or inappropriate content, we want to know. All reports are reviewed manually by our team.

Email us at support@coursebox.ai or visit the Security page for full details on data protection and our AI providers.