AI Use & Transparency
Principles. AI is optional, teacher-controlled, minimal in what it sends, never used to train models on your content, and never makes final decisions about a student.
What AI does in Fikra Smart
Lesson and slide generation, worksheets and quizzes, simulation and interactive-lesson creation, text-to-speech, teacher assistant, student hints, sketch feedback, image reading for maths rendering, and similar tools.
Providers
Requests go from our server to Google (Gemini), Anthropic (Claude) or DeepSeek, as selected by Fikra Smart and listed in Sub-processors. Images and speech use Google. Provider locations may be outside your country; see the Privacy Policy on transfers.
What is sent
- The prompt you or your student wrote, relevant lesson text or images you selected, and system instructions that shape the answer.
- Not sent by design: account email, passwords, rosters, marks, or student identifiers.
- Free text and images you provide may contain personal data if you type it in; do not include students' personal data in prompts or images.
Training and retention
We do not use your content to train or fine-tune any AI model. We use provider API services; providers may retain requests for a limited time for abuse monitoring as set out in their terms. We keep only counts of your AI use, not a copy of your prompts, apart from the content you save in your lessons.
Human oversight
- Teachers must review all AI output before showing it to students.
- AI hints can be turned off per session. AI sketch scoring is a suggestion; the teacher confirms the score. No AI output decides marks, progression or discipline.
Limits and risks
AI can produce wrong, biased, outdated or inappropriate content, and may produce text that resembles existing works. Maths and science answers must be checked. Safety filters reduce but do not remove risk. Report problematic output to support@fikrasmart.com; we will investigate and add safeguards.
Compliance
We follow data-protection law (GDPR, UAE PDPL, COPPA and others) for AI processing, and align our practice with the OECD AI Principles, UNESCO's guidance on AI in education, the transparency expectations of the EU AI Act (users are told when content is AI-generated) and UAE national AI guidance. We do not use AI for emotion recognition, biometric identification or social scoring.