Safeguarding & Child Safety Policy

1. Who this policy applies to

This policy applies to all users of the LaunchPard platform, including scholars (children), parents and guardians, and any future staff, contractors, or partners who have access to the platform or its data. “Scholar” means any child or young person using LaunchPard, regardless of their age or location.

2. Platform design for child safety

LaunchPard is designed from the ground up with child safety in mind.

2.1 No direct communication

There is no messaging, chat, video calling, or any form of direct communication between users on LaunchPard. Scholars cannot contact other scholars, parents, or staff through the platform. There are no social features, comment sections, forums, or user-generated content visible to other users.

2.2 No 1:1 tutoring

LaunchPard does not provide live tutoring, video lessons, or any form of real-time interaction between children and adults. All learning is self-paced and delivered through our adaptive question engine and AI tutor (Tara).

2.3 Parental control by design

Only parents or guardians can create accounts. Scholars cannot sign up independently. Parents control which curriculum their child follows, can view all learning data through the parent dashboard, and can delete their child’s account at any time.

2.4 Age-appropriate content

All questions in our question bank are curriculum-aligned and age-appropriate. Questions are generated to match the scholar’s enrolled curriculum and year level. Our content review process ensures that no violent, sexual, discriminatory, or otherwise inappropriate material enters the platform.

3. AI safety — Tara, our AI tutor

Tara is LaunchPard’s AI tutor, powered by large language models. Tara uses an “Explain It Back” pedagogy where scholars articulate their reasoning after answering questions. The following safeguards are in place:

3.1 Input filtering

All scholar input to Tara passes through a client-side profanity filter that blocks inappropriate language before it reaches the AI. Scholars who attempt to use blocked words receive a gentle prompt to rephrase.

3.2 AI output safety

Tara’s API route uses Meta’s Llama Guard safety model to screen both inputs and outputs. The AI system prompt is carefully constructed to keep Tara focused on the curriculum topic, the specific question being discussed, and age-appropriate language. Tara is instructed never to discuss topics outside the curriculum, never to share personal opinions, and never to ask scholars for personal information.

3.3 Conversation limits

Tara interactions are strictly bounded: one explanation attempt and one follow-up question per quiz question. There is no open-ended conversation. All interactions are stateless — Tara has no memory of previous sessions.

3.4 No personal data collection through AI

Tara does not collect, store, or process personal data beyond what is needed for the immediate interaction (the scholar’s first name and year level, used to calibrate the response). Scholar inputs to Tara are not stored permanently or used for AI training.

4. Data protection

4.1 What we collect

We collect the minimum data necessary to deliver the service:

• Parent: name, email address, region, subscription status
• Scholar: first name (or chosen display name), curriculum, year level, quiz answers, mastery scores, badges earned, coins earned

We do not collect dates of birth, photographs, physical addresses, phone numbers, school names, or any biometric data from scholars.

4.2 How we use it

Scholar data is used solely to:

• Deliver adaptive learning (selecting appropriate questions and tracking mastery)
• Display progress to the scholar and their parent through dashboards
• Power gamification features (leaderboards use display names only)

We do not sell, share, or transfer scholar data to any third party for marketing, advertising, or profiling purposes.

4.3 Where it’s stored

All data is stored in Supabase (hosted on AWS infrastructure) with row-level security policies enforced at the database level. Data is encrypted in transit (TLS 1.2+) and at rest. Our infrastructure is hosted on Vercel (for the application) and Supabase (for the database), both of which maintain SOC 2 Type II compliance.

4.4 Data retention and deletion

Parents can request deletion of their child’s data at any time by emailing privacy@launchpard.com. Upon receiving a verified deletion request, we will permanently delete all scholar data within 30 days. When a parent deletes a scholar profile through the dashboard, associated quiz results, mastery records, and learning path data are permanently removed.

4.5 Applicable regulations

We comply with:

• UK: UK GDPR, Data Protection Act 2018, Age Appropriate Design Code (Children’s Code)
• Nigeria: Nigeria Data Protection Regulation (NDPR), Child Rights Act 2003
• Canada: Personal Information Protection and Electronic Documents Act (PIPEDA)
• Australia: Privacy Act 1988, Australian Privacy Principles

5. Leaderboards and social features

Our leaderboards display only the scholar’s chosen display name (first name or codename) and their score. No photographs, locations, ages, or identifying information are shown. Leaderboards are curriculum-scoped — scholars only see others in the same curriculum and year level. There is no way for scholars to contact, follow, friend, or interact with other scholars through the platform.

6. Third-party services

LaunchPard uses the following third-party services that may process data:

• Supabase — database and authentication (stores all user data)
• Vercel — application hosting (processes HTTP requests)
• OpenRouter / AI providers — processes Tara AI interactions (scholar first name and question context only; no persistent storage)
• Brevo — transactional email to parents only (never to scholars directly)
• Google Analytics — anonymised usage analytics (no personally identifiable information)

We do not use any advertising networks, social media trackers, or third-party profiling services. Google Analytics is configured without collecting personal data and with IP anonymisation enabled.

7. Reporting concerns

If you have any safeguarding concerns about a child using LaunchPard, or if you believe the platform has been used inappropriately, please contact us immediately:

If you believe a child is in immediate danger, please contact your local emergency services or child protection agency directly. In the UK, you can also contact the NSPCC helpline on 0808 800 5000 or Childline on 0800 1111. In Nigeria, contact the National Agency for the Prohibition of Trafficking in Persons (NAPTIP) helpline on 07030000203.

8. Content moderation

All questions in the LaunchPard question bank are either sourced from established curriculum frameworks or generated by AI models and reviewed for appropriateness. Our content pipeline includes automated quality checks for offensive language, factual accuracy, and age-appropriateness. Questions that fail these checks are rejected before entering the live question bank.

Parents who encounter inappropriate content can report it through the parent dashboard or by emailing safeguarding@launchpard.com. Reported content is reviewed and, if confirmed, removed within 24 hours.

9. Staff and access controls

Access to scholar data is restricted to essential personnel only. Database access requires multi-factor authentication. All administrative actions are logged. We do not employ individuals who work directly with children — all learning is delivered through the software platform.

10. Policy review

This policy is reviewed at least annually and updated whenever there are material changes to the platform, applicable regulations, or our understanding of risks to children. The date at the top of this page reflects the most recent review.

11. Contact

LaunchPard Technologies Ltd
Email: safeguarding@launchpard.com
Data protection enquiries: privacy@launchpard.com