Legal
AI Transparency Statement
Last updated: 15 July 2026 · v1.0
Oiva is an AI-native platform for Australian early childhood education and care. We believe the sector deserves AI that is safe, sovereign and accountable. This statement explains, in plain language, how AI is used in Oiva, the rules it operates under, and the rights you have. It reflects our two commitments: AI that asks before it acts, and AI that stays in Australia.
1. Our AI principles
Human in the loop, always.
AI in Oiva drafts, summarises, estimates and flags. It never finalises anything that matters. Every AI output that could affect a child's record, a family's subsidy, a payment or a compliance obligation must be reviewed and approved by a qualified person before it takes effect. There is no auto-approve path in the architecture.
Sovereign by design.
All AI processing runs on infrastructure physically located in Australia. No personal information leaves Australia for AI processing. This is enforced technically, not just by policy.
Privacy by design.
We minimise personal information in AI prompts and structurally exclude financial identifiers (such as tax file numbers, bank details and Customer Reference Numbers) from AI processing, with automated tests verifying the exclusion.
No training on customer data.
Information entered into Oiva is never used to train or fine-tune AI models.
No biometrics.
Oiva does not perform facial recognition or any other biometric identification or processing.
Transparency.
When you interact with AI (for example, an assistant), it is identified as AI. AI-drafted content is labelled as AI-drafted until a person approves it.
Accountability and auditability.
AI interactions are logged with redaction and retention controls. Every human approval of an AI-assisted action is recorded, creating a complete audit trail.
2. Where AI is used in Oiva
- Drafting assistance for educators (for example, observations and documentation), always subject to educator review.
- Surfacing relevant compliance information and flagging records that may need human attention.
- Estimating fees and subsidy outcomes to support conversations with families. These are estimates only, confirmed by government systems and human review.
- Assistants and chat interfaces, clearly identified as AI.
3. Where AI is not used
- AI does not decide eligibility, entitlements, enrolment outcomes or payments.
- AI does not take adverse action against any person.
- AI does not process children’s images biometrically or profile children.
- AI does not operate on financial identifiers.
4. Alignment with Australian frameworks
Our AI governance is aligned with the Australian Government’s Guidance for AI Adoption (the six essential practices), the OAIC’s guidance on privacy and the use of AI, and Australia’s AI Ethics Principles.
| AI6 practice | How Oiva meets it |
|---|---|
| 1. Accountability | Named executive accountability for AI governance; documented decision authority; AI register maintained. |
| 2. Risk management | Risk assessments and privacy impact assessments before AI features ship; red-team testing including prompt injection resistance; fail-closed release gates. |
| 3. Testing and monitoring | Offline evaluation sets, continuous evaluation in production, drift monitoring, and defined rollback triggers when quality or safety regresses. |
| 4. Transparency | AI features identified to users; AI-drafted content labelled; this public statement; privacy policy disclosure of AI and automated processing. |
| 5. Privacy and data governance | Prompt minimisation and redaction; structural exclusion of financial identifiers; Australian-only processing; no training on customer data; alignment with the APPs and OAIC AI guidance. |
| 6. Continuous improvement | Post-incident reviews, evaluation metrics tracked over time, and versioned prompts and policies under change control. |
5. Your rights and how to raise concerns
If you believe an AI-assisted output in Oiva is wrong, you can ask the service using Oiva to review and correct it. A person is always able to override AI-drafted content. You can also contact us directly at hello@oiva.com.au.
Privacy-related concerns are handled under our Privacy Policy, including your rights of access, correction and complaint to the OAIC.
6. Review
This statement is reviewed at least annually and whenever we materially change how AI is used in Oiva.
Related legal documents