School checklist before uploading real pupil data

This teacher-facing checklist helps a school, trust, DPO, IT lead, or department record the review needed before Feedback Flow is used with identifiable pupil work. It supports DPIA evidence gathering, but it is not legal advice and does not replace the school's own governance process.

Review details

School or trust

Review date

Teacher or project owner

DPO / IT approval owner

Intended use and pupil data involved

Keep this page with the school's DPIA or supplier review record. The form does not upload or save these notes; printing and text export happen in the browser.

Lawful basis and transparency

Controller decision

The school has recorded its lawful basis for using Feedback Flow with pupil work.For many UK school assessment workflows this may be public task, legitimate interests, contract, or another basis depending on the institution. Do not assume consent is the default.
Privacy information for pupils, parents, and staff has been checked.Confirm whether the school privacy notice, AI policy, acceptable-use policy, or classroom notices need updating before rollout.
Special-category or highly sensitive data routes are documented.If pupil work may contain health, SEND, safeguarding, ethnicity, religion, or other sensitive content, record the Article 9 condition or prohibit upload for those cases. PP/SEND indicators should be approved as limited fields, not free-text case records.

Evidence or conditions

DPO and IT approval

Before rollout

The DPO or data protection lead has reviewed the intended use.Attach the DPIA decision, risk rating, and any agreed safeguards or restrictions.
The IT, cyber, or information security lead has reviewed supplier and access controls.Include authentication, browser access, API key handling, hosting, sub-processors, and staff onboarding/offboarding.
The Data Processing Agreement and processor/sub-processor list have been reviewed.Confirm controller/processor roles, international transfers, retention, audit support, deletion, and breach notification routes.

Approval evidence

Data minimisation

Upload discipline

Staff know which identifiers to remove before upload.Avoid unnecessary names, candidate numbers, UCI numbers, email addresses, phone numbers, addresses, and original filenames.
Staff know which pupil work must not be uploaded without extra approval.Minimal PP/SEND indicators may be used only if approved. EHCP documents, SEND case files, safeguarding disclosures, medical details, behaviour logs, family circumstances, and high-risk personal narratives need a clear separate school rule.
Only necessary mark schemes, context, rubrics, and class details are included.Use the smallest useful dataset for the teaching purpose and avoid uploading whole folders when a single response set is enough.

Minimisation rule

Gemini paid API review

AI processor check

The school has reviewed the paid Gemini API data-use position.Google Cloud terms state that customer data is not used to train or fine-tune models without permission; the school should verify the current terms during procurement.
Logging, caching, abuse monitoring, and retention limits are understood.Record any configuration decisions, including whether request/response logging is disabled and whether features with additional retention are used.
Staff understand that AI processing is server-side and not a free consumer chatbot workflow.Do not paste pupil work into unrelated public AI tools to work around school controls.

Gemini review notes

Retention and deletion

Lifecycle control

The school has set a retention period for uploaded pupil work and AI outputs.Align the period with assessment, appeal, moderation, and local record-management needs.
Deletion and account closure routes are known before upload.Confirm who can request deletion, when exports are needed first, and whether school-held local copies also need deletion.
Data export needs have been checked.If the school needs a record for moderation, dispute handling, or internal audit, decide what must be exported before deletion.

Retention schedule

Teacher review and professional judgement

Human control

Teachers must review transcriptions, marks, feedback, and analysis before use.AI output is draft support. It must not be treated as final professional judgement without teacher checking and approval.
The school has a route for correcting or rejecting poor AI output.Teachers should be encouraged to edit, regenerate, ignore, or escalate outputs that are inaccurate, biased, unfair, or inappropriate.
AI is not used as the sole basis for grades, setting, interventions, SEND, behaviour, or safeguarding decisions.Those decisions require local policy, professional judgement, and appropriate evidence.

Teacher review rule

Assessment and NEA caution

Integrity check

Live coursework, controlled assessment, NEA, and exam-board evidence are prohibited unless explicitly approved.Check centre policy, awarding organisation rules, JCQ or exam-board guidance, and any moderation or appeals requirements before use.
Approved use is limited to appropriate formative or teacher-reviewed workflows.Record whether use is allowed for practice essays, mock feedback, departmental moderation support, or analysis of already-marked work.
Student AI-use boundaries are separate and clear.This checklist covers teacher use of Feedback Flow. It does not permit pupils to submit AI-generated work as their own.

Assessment boundary

Role and access controls

Least privilege

Only authorised staff can upload or view pupil work.Confirm whether access is limited to named teachers, department leads, administrators, or school-managed accounts.
Shared accounts and unmanaged credentials are prohibited.Use individual accounts so access, deletion, support, and incident response can be managed properly.
Staff leaver and role-change processes include Feedback Flow access.Decide who owns class data if a staff member leaves and when access should be removed.

Access control evidence

Incident contact and escalation

Response route

Staff know the local route for privacy, safeguarding, cyber, or assessment-integrity concerns.Include the school DPO, IT helpdesk, safeguarding lead, exams officer, or senior owner as appropriate.
Feedback Flow contact details are recorded for data protection and procurement questions.Use [email protected] for privacy, school procurement, security, or account deletion questions.
The school has considered breach-notification thresholds and evidence capture.Record what staff should preserve if pupil data is uploaded incorrectly or an output creates a risk.

Incident contacts

Evidence export and print

Audit pack

Relevant documents are attached or linked in the school's DPIA record.Include the AI and Data Use page, Privacy Policy, DPA, supplier answers, and any local policy decision.
The completed checklist has been printed, saved as PDF, or downloaded as an evidence log.Use the buttons at the top of this page. The text export includes checked items and evidence notes entered on this page.
A review date has been set.Repeat the review after major product, provider, policy, legal, assessment, or school governance changes.

Decision

Decision owner

Next review date

Conditions or restrictions

Decision gates

No real pupil data uploaded until DPO/IT approval is recorded.
Upload rule removes unnecessary identifiers and sensitive content.
Teacher review remains mandatory for every AI output.
Evidence pack saved with the school DPIA record.

Documents to attach

AI and Data Use
Privacy and Governance Summary
Privacy Policy
Data Processing Agreement
Local school AI, data protection, safeguarding, and assessment policies

Useful external references

Schools should verify current guidance as part of their own review.