Safeguarding Policy

1. Purpose

This policy sets out how safeguarding responsibility works on the Tutinama platform, what the platform does and does not do in relation to pupil welfare, and how information that may be safeguarding-relevant is handled.

2. Who is responsible for safeguarding

The tutor is responsible for the safeguarding of their pupils. This includes their own professional judgement, any DBS or vetting checks appropriate to their practice, and deciding how to act on any concern about a pupil's welfare — whether that concern arises in person, through information on Tutinama, or anywhere else.

Tutinama is a technology platform. We do not employ tutors, supervise tutoring sessions, hold a designated safeguarding lead, or carry out DBS checks on tutors using the platform. Tutors are independently responsible for their own vetting, qualifications, and professional standing, and are expected to hold any DBS check or professional accreditation appropriate to working with the age group they tutor, independently of Tutinama.

3. What the platform does to support the tutor's role

Tutinama gives tutors tools relevant to pupil wellbeing:

• A wellbeing notes field, encrypted at rest, visible only to the tutor who entered it. These notes are never shown to the AI Tutor Assistant and never used to shape a pupil's practice sessions.
• A tiered safeguarding detection system built into the AI Tutor Assistant itself, described in full in Section 4, which can raise an immediate alert to the tutor (and, where applicable, the guardian) during a live session rather than waiting for a routine summary.
• Flags generated from ordinary session activity, surfaced to the tutor alongside the session summary, intended to draw attention to something in a pupil's practice session that may be worth a closer look.

These tools support the tutor in noticing something and, for the most serious cases, ensure a human is alerted promptly — they do not constitute Tutinama itself assessing, triaging, or acting on a safeguarding concern. That responsibility remains with the tutor at all times.

4. Concerning content raised directly to the AI Tutor Assistant

The AI Tutor Assistant uses a three-tier system to recognise and respond to concerning content raised by a pupil during a session:

• Tier 1 — explicit disclosures of self-harm, suicide, abuse, or danger. The AI is instructed, through a system-level rule that cannot be overridden by any tutor configuration, to respond with warmth and to signpost appropriate crisis support. This triggers a structural, system-enforced alert (rather than relying on the AI reliably including a hidden marker in its own text, which testing showed was unreliable in genuinely emotional moments) that immediately logs the event, marks the session as a safeguarding alert, and sends an automated email to the tutor, and to the guardian where the pupil is under 18 and a guardian email is on file. That email states who, when, and the general subject only — it does not contain a transcript or quote the pupil's disclosure, keeping the pupil's own words private while still ensuring a human is alerted promptly.
• Tier 2 — lower-severity distress (for example, a pupil saying they're struggling to cope). The AI pauses the academic content, acknowledges what the pupil has said, and checks in once, without raising an alert to the tutor.
• Tier 3 — ordinary exam stress or frustration. The AI briefly acknowledges this and continues, with no special action taken.

Administrators also see a platform-wide count of safeguarding alerts on the admin dashboard, for visibility. There is no case-management system beyond this and no further content logging — by design, follow-up from a Tier 1 alert is a human process between tutor, guardian, and pupil, not something Tutinama manages or mediates.

5. Data handling for safeguarding-relevant information

Wellbeing notes and anything flagged as a safeguarding concern are retained on the same basis as standard session data — 12 months after account closure, then deleted — and access is restricted to the tutor who recorded it.

6. Escalation

Where a tutor identifies a safeguarding concern, escalation (for example, contacting a parent, an appropriate professional, or the relevant authorities) is the tutor's responsibility, using their own professional judgement. Tutinama does not receive, triage, or act on safeguarding disclosures, and does not act as an intermediary between a tutor and external services.

7. Complaints or concerns about a tutor's conduct

If a parent, guardian, or pupil has a concern about a tutor's conduct rather than a pupil's welfare, they can raise this with Tutinama directly at jamie.mata981@gmail.com. We will take this seriously and, where appropriate, may restrict a tutor's access to the platform pending review.

8. Review

This policy will be reviewed periodically as the platform develops.