Transparency

AI Use Disclosure

Last updated: 18 April 2026

1. Why this Disclosure Exists

Mirame Australia Pty Ltd (ABN 97 696 515 616) believes users should know when and how automated systems influence the service they receive. This Disclosure explains exactly what "AI" means in Mirame, what the algorithm does, what it does not do, and what you can do about it.

We publish this Disclosure in line with:

2. What the Algorithm Does

Mirame's colour season assignment is produced by a deterministic, rule-based scoring algorithm. In plain English:

Technically this is a rule-based algorithm (sometimes called "classical AI" or "symbolic AI"). It is not generative AI. It is not a large language model. It is not machine learning in the sense of a neural network trained on data.

We disclose it as "AI" because users commonly refer to any automated decision as "AI", and because ACL s18 requires us to describe the product accurately rather than rely on a narrow technical distinction.

3. What the Algorithm Does Not Do

4. What This Means for You

4.1 Right to explanation

You can ask us at any time to explain, in plain English, why the algorithm assigned you the season it did. We will walk you through which of your answers pushed the score toward that season.

4.2 Right to human review

If you believe the algorithm has given you the wrong season, you can request a human review before or in addition to requesting a refund. To request a review, email hello@mirame.au with:

A human reviewer will re-check your answers, offer a second opinion, and — if warranted — adjust your assigned season and reissue the paid reading content at no cost.

Response window: 14 days.

4.3 Right not to submit sensitive data

Human review is entirely optional. You are not required to send a photograph of yourself. If you do voluntarily submit a selfie, we will use it only for the review, we will not store it beyond the time needed to resolve the review, and we will not use it for any other purpose. See the Privacy Policy for more detail.

4.4 Right to human contact

If you'd rather speak with a person than use the automated quiz at all, we're a small team — email hello@mirame.au and we'll help.

5. Limits of the Service

Mirame is guidance, not prescription. In particular:

6. Governance and Oversight

The algorithm, scoring weights, content, and retailer matches are reviewed and approved by Mirame's product team before release. Changes to the algorithm or content are version-controlled and logged. If you have questions about the methodology, see the Methodology Acknowledgement page or contact us.

The Service is subject to regular security review under the ACSC Essential Eight and OWASP frameworks by our security function (Sentinel), including specific review of AI/ML components per the OWASP Top 10 for Large Language Model Applications guidance where it becomes relevant (e.g. if generative AI is added to the Service in the future). A new version of this Disclosure will accompany any such addition.

7. Changes to this Disclosure

If we materially change what AI does in Mirame — for example, if we add a generative element, integrate a large language model, or move to adaptive personalisation — we will update this Disclosure, publish a changelog, and notify active paid reading users by email before the change goes live.

8. Contact

This is compliance tooling, not legal advice. Blake's lawyer finalises before launch. Sentinel to review AI-specific language for alignment with the Voluntary AI Safety Standard and OWASP Top 10 for LLM Applications (where future versions introduce generative components).