Australia · School assessment
How can Australian schools review writing in the AI era?
Australian schools can strengthen take-home writing by setting clear AI-use rules, collecting progressive evidence and keeping decisions with teachers. Loam provides an Australian-operated writing workspace with Sydney as its Australian data region, no camera or screen recording, and a reviewable process record. It supports school policy; it does not replace it.
- Primary question
- How can an Australian school use process evidence while preserving human judgement and privacy?
- Search intent
- Evaluate an Australian-hosted writing-process workflow against school assessment, privacy and procurement needs.
What does Australian guidance ask schools to do?
Define permitted AI use, preserve human judgement and design assessment so students can demonstrate learning in more than one way. The Australian Framework for Generative AI in Schools addresses responsible and ethical use, including privacy, fairness, transparency, accountability and teaching and learning.
In NSW, NESA says schools remain responsible for their AI policies and specific assessment advice. Unapproved AI use can breach academic integrity, while approved use must be acknowledged under the applicable rules. NESA also points to varied assessment tasks, progressive evidence and student explanation.
These sources do not prescribe Loam or endorse a particular product. They establish the policy context a school should test any product against.
Where does an Australian Loam workspace store its data?
A workspace that selects the Australia region stores its Loam service data in AWS Sydney. Loam's current Privacy Policy and DPA say stored work, evidence records, account data and encrypted backups remain in the selected region.
Regional storage is not a claim that every supporting service operates only in Australia. Google or Microsoft sign-in, Stripe billing and other subprocessors have separately disclosed processing locations and data scopes. Schools should review the subprocessor register and the applicable agreement rather than relying on a residency slogan.
What does Loam observe about a student?
Loam records the document process, not a video or biometric profile of the student. Its published privacy commitments exclude camera and microphone access, screenshots, screen recording, monitoring other tabs or applications, and keystroke-dynamics identity profiles.
The system does record document changes, revisions, timing and evidence checkpoints inside the Loam editor. That is personal information in a school context and still requires a defined purpose, clear notice, controlled access, suitable retention and an assessment process that accounts for accessibility.
What should an Australian school test before a pilot?
Test the evidence boundary, the data flow and the human review together. A hosting region is useful, but it does not answer whether the task is valid, the collection is proportionate or the resulting record is interpreted fairly.
Questions for an Australian school review
- Where will stored student work and evidence remain?
- Selected region
- For an Australian workspace, Loam's terms identify AWS Sydney for stored work, evidence, account data and encrypted backups. Review separately listed identity and billing services too.
- What observation does the product perform?
- Document activity
- Loam records work inside its editor. It says it does not use a camera, microphone, screenshots, screen recording, other-tab monitoring or behavioural biometrics.
- Who makes an integrity decision?
- The school
- Loam exposes a record for review. It does not determine authorship, AI use or misconduct, and government guidance remains subject to each school's policy.
- Can the workflow accommodate different needs?
- Must be planned
- Schools should decide suitable adjustments, permitted assistance and alternative evidence before the task rather than treating one capture pattern as universal.
- Can reviewers inspect the evidence boundary?
- Yes
- Loam publishes a synthetic demonstration, local browser verifier, privacy documents, DPA, subprocessor list and explicit cannot-prove statements.
| Question | Answer | Evidence boundary |
|---|---|---|
| Where will stored student work and evidence remain? | Selected region | For an Australian workspace, Loam's terms identify AWS Sydney for stored work, evidence, account data and encrypted backups. Review separately listed identity and billing services too. |
| What observation does the product perform? | Document activity | Loam records work inside its editor. It says it does not use a camera, microphone, screenshots, screen recording, other-tab monitoring or behavioural biometrics. |
| Who makes an integrity decision? | The school | Loam exposes a record for review. It does not determine authorship, AI use or misconduct, and government guidance remains subject to each school's policy. |
| Can the workflow accommodate different needs? | Must be planned | Schools should decide suitable adjustments, permitted assistance and alternative evidence before the task rather than treating one capture pattern as universal. |
| Can reviewers inspect the evidence boundary? | Yes | Loam publishes a synthetic demonstration, local browser verifier, privacy documents, DPA, subprocessor list and explicit cannot-prove statements. |
Claim boundary
What can Loam support, and what remains the school's responsibility?
Can show
- Captured edits, revisions, timing and checkpoints inside the Loam editor.
- Whether a supported evidence record passes Loam's published integrity checks.
- The selected data region and the current product privacy and processing commitments.
- A replay and evidence record a teacher can examine during a documented review.
Cannot prove
- Who was physically at the keyboard or where every idea originated.
- That no unrecorded assistance, collaboration, device or external tool was used.
- That a student breached school or NESA rules.
- That one workflow is appropriate for every task, student, disability or adjustment.
Loam is evidence infrastructure for a teacher-led process. It is not a regulator, government-endorsed product, identity system, AI-use detector or automated academic-integrity decision maker.
How should a school approach Australian privacy obligations?
Start with the school's applicable law and data-governance responsibilities, then test Loam's commitments against that standard. The OAIC explains that the Australian Privacy Principles cover collection, use, disclosure, governance, security, access and correction for entities within their scope. Public schools can also be subject to State or Territory regimes.
Difinity's Privacy Policy says it voluntarily applies the APP commitments regardless of whether a small-business exemption might otherwise apply. That statement is a product commitment, not a certification or a substitute for the school's own review.
Primary source: Office of the Australian Information Commissioner, Australian Privacy Principles.
Source ledger
Which primary sources support this page?
- 01Australian Department of Education — Framework for Generative AI in Schools
National framework for responsible and ethical generative-AI use across Australian school education. Last modified 17 June 2025.
- 02NESA — Use of Artificial Intelligence by students
NSW guidance on academic honesty, acknowledgement, school policy, varied assessment and fair application of plagiarism controls.
- 03NESA — 2026 HSC Rules and Procedures
Current HSC guidance covering acknowledgement, unauthorised AI use, progressive evidence and student explanation.
- 04OAIC — Australian Privacy Principles
Primary regulator overview of the principles governing collection, use, disclosure, governance, security, access and correction of personal information.
- 05Loam — Privacy Policy
Current product commitments for capture boundaries, selected-region storage, third-party services, access and retention.
- 06Loam — Data Processing Agreement
Institutional terms covering processing instructions, regional storage, security measures, subprocessors and deletion.