Qwizflow is committed to the responsible and ethical use of Artificial Intelligence in education. Our AI practices are guided by five core principles:
AI is used exclusively to enhance learning outcomes. Every AI feature is designed to help students understand, practice, and retain knowledge.
All AI-generated content is clearly labelled so students always know when they are interacting with AI. We never disguise AI outputs as human-created content.
Student data is never used to train AI models. Before anything reaches an AI model we mask contact details and labelled names, and a learner's account name, email address and date of birth are never included (see PII Scrubbing below for what this can and cannot catch). AI consent is saved with the account, purpose by purpose; nothing is switched on until it is chosen, and it can be changed or withdrawn at any time in the AI Consent Centre.
Our AI features are designed to be accessible and equitable for all students regardless of background, ability, or learning style.
Strict content safety filters block harmful, explicit, or dangerous material. Prompt injection protections prevent misuse of AI features.
Qwizflow uses AI to power the following educational features, each of which requires explicit user consent before activation:
Generates personalised quizzes from uploaded study materials and curriculum topics.
Guides students through concepts using questioning techniques rather than giving direct answers.
Provides step-by-step hints and scaffolding for homework problems without revealing answers directly.
Creates structured study guides from uploaded documents and learning paths.
Adjusts difficulty and topic sequencing based on quiz performance.
AI-powered educational simulations for critical thinking and creative writing practice.
Collaborative creative writing with AI, where students and AI take turns building a narrative.
Generates personalised learning companion avatars and accessories.
Provides detailed explanations for quiz answers to support understanding.
Qwizflow explicitly prohibits the following uses of AI:
To adapt help, Qwizflow can notice how a learner is getting on with a topic — for example retries, hints and time spent on a question — but only when that learner’s AI consent allows it. Teachers see these signals only as class-wide patterns drawn from at least five learners, never for one student.
Qwizflow uses the following AI providers under strict data processing agreements:
Content generation (quizzes, study guides, tutoring), pictures and Studio Voice, using Google's current Gemini models; the read-aloud voice uses Google Cloud Text-to-Speech. Private API calls only — student data is never used for model training.
Reads the text in uploaded photos of study material (PDFs are read on our own servers). It is text recognition, not a generative AI model, and it receives no prompt — only the image. Never used for learners under 13, who cannot upload files.
Before any content is sent to AI providers, a scrubbing layer masks email addresses and phone numbers, strips known prompt-injection patterns, caps prompt length, and redacts names where they appear behind an explicit label. Account names, email addresses and dates of birth are never included in a prompt.
We used to describe this as “the AI never processes student identities”. That overstated it, so here is the accurate version: the scrubber is pattern-based, so it cannot reliably catch personal details a learner types into free text themselves. Under-13 Kids Mode is tap-only, which removes that path for children entirely. Prompts do carry learning context — topics, strengths and misconceptions — because that is what makes help personalised, and that is disclosed in our Privacy Policy.
Student data and uploaded documents are processed via private API calls and are NEVER used to train public AI models by Google or Microsoft. Training data opt-out is enabled by default.
Qwizflow maintains an AI Ethics Officer responsible for:
These responsibilities sit with Qwizflow’s founder, who is the named accountable person for AI ethics and data protection while the team is this size. We would rather say that plainly than imply a larger governance function than exists. The role can be reached through the in-app feedback portal, and all concerns are reviewed within 5 business days.
If you have concerns about how AI is used in Qwizflow, or if you encounter inappropriate AI-generated content, you can: