Transparency
AI transparency
Where we use AI
AI is used to generate and check course content, provide the AI Trainer and guest roleplay, generate lesson practice quizzes, produce practice scorecards and suggest learning pathways. In the trainer and roleplay you interact with AI, not a human. Multiple-choice exam and diploma unit quiz results are computed automatically from answer keys, rather than an AI judging the answer.
Providers and processing
The current text-based AI features use OpenAI through the Lovable AI Gateway. Lovable Cloud stores account and learning records; Cloudflare delivers the website. Requests may include the conversation, relevant course context or profile details needed for the feature. Do not put confidential guest information or sensitive personal data into a prompt. Diploma submissions are stored for assessment and are not sent by the trainer to the model.
Processing may take place outside the UK, including in the EEA and the United States. We do not claim that all processing stays in one country. The precise processing location and applicable transfer safeguards must be confirmed for the service configuration; our Privacy Policy explains international transfers.
Limits of AI
AI output can be wrong, biased or incomplete. It is not professional, legal, medical or financial advice. Course checks and practice feedback cannot guarantee accuracy, safety or a particular career outcome.
Data and optional improvement
We do not use learner conversations or submissions to train third-party models or send them to third-party training services. Our AI requests use non-stored inference, but this does not by itself prove that providers keep no operational or safety logs. Contractual provider retention and training protections require confirmation. Use to improve our own academy system requires a separate opt-in on your Account page, off by default. No own-system improvement programme is currently running. Withdrawing permission does not affect learning or assessment.
Automated results and human review
Exams and unit quizzes produce automatic results. You can request a human review of any automated result, at no cost, within 20 working days of receiving it. Use Contact, topic Assessment appeal. A human, not AI, decides appeals and malpractice cases. This route supports your rights under UK GDPR Article 22, where applicable, and the KVKK right to object to outcomes based solely on automated processing.
EU learners and oversight
For EU learners, we treat AI used to evaluate learning outcomes as high risk under the EU AI Act. Human oversight, assessment logging, question accuracy monitoring and these transparency notices form our approach. This is not a claim of certification or a completed legal compliance assessment. Diploma results, certificate revocations and disputed academic decisions require human approval.
Report a problem
Use Report this answer beside a completed AI Trainer reply and explain the problem. The answer and your reason are stored privately for an administrator to review. Reporting is not a substitute for an assessment appeal; use Contact for a disputed result.