Privacy Policy

1.1 What UNCAPT builds

UNCAPT is the creator of the ELI Platform, an operating system for expert reasoning that captures, governs, scales, and compounds domain expertise. Healthcare is our first domain; the platform is domain-agnostic. Our current products are MIA (mental health triage), ARI / Insight (research intelligence), MATILDA (clinical decision support), CARA (aged care reablement) and NADIA.

UNCAPT is an APP entity under the Privacy Act. Where our products are deployed within a Partner Organisation's platform, both UNCAPT and the Partner Organisation hold obligations to individuals under Australian privacy law.

Partner Organisation

The primary entity that holds the direct relationship with the individual (also referred to as “Partner” or “Customer” in UNCAPT's commercial agreements). Responsible for obtaining consent, managing the primary relationship, and meeting obligations under the Privacy Act and applicable health records legislation.

UNCAPT

As an entity that “holds” and processes information in the course of providing the platform, UNCAPT has direct obligations under the APPs, including security (APP 11), data quality (APP 10), and use and disclosure (APPs 6 and 7), regardless of contractual arrangements with the Partner Organisation.

We do not rely on our role as a technology provider to limit our direct responsibilities to individuals whose data we process.

UNCAPT interacts with three distinct categories of individuals.

Direct patients and end users

Individuals using MIA for mental health triage. UNCAPT collects health information and triage data directly from these individuals, with their express informed consent, at the point of use.

Expert users and clinicians

Clinicians and domain experts using the ELI Platform to review, edit, validate, or scale expert reasoning. UNCAPT collects professional details and activity records to maintain accountability and the integrity of clinical reasoning outputs.

Programme users

Individuals who use UNCAPT products through a Partner Organisation's programme in a non-clinical capacity, for example seeking guidance or information rather than assessment or care. Programme users are not patients. We collect the content of their interactions, which may include information about other people; users are asked not to include identifying details, and names are removed before storage.

This distinction determines how consent is obtained, what information is collected, and how breach notifications are communicated, as set out in the relevant sections below.

2.1 Unsolicited information (APP 4)

If we receive personal information we did not solicit, we will assess whether it could have been collected under APP 3. If not, we will destroy or de-identify it as soon as practicable, provided it is lawful and reasonable to do so.

2.2 Information you provide

• Contact information: Name, email address, organisation, and role when you request access, enquire about partnerships, or submit an enquiry form.
• Direct patient / end-user data (via MIA): Health information, triage responses, and session data collected directly from individuals using MIA, based on their express informed consent at the point of triage.
• Expert / clinician data (via ELI Platform): Professional details (name, role, credentials) and audit logs from clinicians and domain experts using the ELI Platform to review, edit, or scale clinical reasoning. This information is collected to ensure accountability and to maintain the integrity of clinical reasoning edits.
• Clinical or domain data (via Partner Organisations): When UNCAPT products are deployed within a Partner Organisation's platform, de-identified or identifiable data may be processed as directed by the Partner Organisation, subject to their privacy obligations and consent processes.
• Programme user data (via Partner Organisation programmes): Individuals who use UNCAPT products through a Partner Organisation's programme in a non-clinical capacity, for example seeking guidance or information rather than assessment or care. Programme users are not patients. We collect the content of their interactions, which may include information about other people; users are asked not to include identifying details, and names are removed before storage.
• Research participant data: If you participate in a research program involving UNCAPT products, data collection is governed by the relevant ethics approval and separate participant consent forms.

2.3 Information collected automatically

• Usage data: Aggregated, de-identified interaction patterns used to improve platform performance and safety.
• Technical data: Browser type, IP address, and device information when you visit our website.

3.1 Direct marketing and communications (APP 7)

We may use contact information provided by domain experts, Partner Organisations, and trial participants to send relevant updates about UNCAPT, including platform updates, research findings, and partnership opportunities.

We will only do so where you have provided your contact information directly to us and would reasonably expect to receive such communications based on your interaction with us.

Opt-out: Every communication we send includes a clear way to opt out of future communications. You can also opt out at any time by contacting info@uncapt.com. We will process opt-out requests within 5 business days.

We do not use patient or subject health information for direct marketing purposes under any circumstances.

5.1 Your data is not used to train AI models

Customer data and data derived from customer use, including session content, expert corrections, outputs, and de-identified versions of that data, are not used to train or fine-tune AI models, including stage verifiers. Your data remains yours. We process data for the purposes described in this policy, including providing, securing, and supporting the service and the separate use of de-identified reasoning described below.

5.2 Anonymised clinical reasoning

UNCAPT products generate structured reasoning outputs (e.g., assessment logic, care plan rationale, scoring pathways) as part of normal operation. We retain de-identified versions of these reasoning outputs to enable continuous improvement of clinical knowledge bases.

This de-identified reasoning may be reviewed and refined by qualified domain experts (including research partners) as part of the Expert-Led Iteration (ELI) cycle. This process ensures our products' reasoning remains accurate, current, and aligned with best practice.

5.3 De-identification standard

We apply de-identification processes consistent with the OAIC's De-identification and the Privacy Act guidance and the De-Identification Decision-Making Framework published by the OAIC and CSIRO's Data61. Our process includes:

• Removal of all direct identifiers (names, dates of birth, contact details, healthcare identifiers)
• Removal or generalisation of indirect identifiers that could contribute to re-identification
• Assessment of re-identification risk, including the “motivated intruder” test
• Ongoing review of re-identification risk as datasets grow

5.4 Sensitive data and re-identification risk

We acknowledge that health and clinical data carries heightened re-identification risk. To mitigate this, our de-identification of reasoning outputs specifically:

• Strips all session-specific narrative content, retaining only structured reasoning pathways, not underlying patient narratives
• Generalises demographic and contextual information to broad categories
• Does not retain unstructured session transcripts or free-text notes in any de-identified dataset used for knowledge-base improvement. De-identified interaction records may be retained separately for safety audit, subject to the retention periods in Section 9.
• Subjects de-identified datasets to periodic re-identification risk assessments

If at any point we determine that de-identified data carries a non-trivial re-identification risk, we treat that data as personal information and apply the full protections of the APPs.

5.5 Transparency and explainability

UNCAPT products are designed to be explainable. Where a product generates a recommendation, it provides the reasoning pathway so that a clinician or domain expert can understand why a suggestion was made and can review, edit, or override the output. UNCAPT products are decision-support tools: they suggest, experts decide.

We are committed to the principles set out in the OAIC's guidance on privacy and AI, including that individuals should be informed when AI is used to process their information, and that human oversight must be maintained for consequential decisions.

8.1 Data quality (APP 10)

We take reasonable steps to ensure that personal information we hold is accurate, up-to-date, complete, and relevant. For clinical data, accuracy is critical to safety. Clinicians and domain experts can review and correct all product outputs before they are relied upon for decisions.

8.2 Cross-border data transfers (APP 8)

Current clinical deployments and private inference are processed and stored within Australia. Non-clinical account, contact, security and website analytics data is processed in the locations listed in Section 7. Any additional clinical deployment region must be specified in the applicable customer agreement. Where personal information is disclosed overseas, we apply contractual safeguards consistent with our obligations under APP 8.

9.1 Clinical data

Clinical records, including directly collected MIA health and session records where health-records law applies, are subject to the applicable statutory minimums. For records held on behalf of a Partner Organisation, its lawful retention instructions also apply. Relevant requirements for covered health service providers or record keepers include, subject to statutory exceptions:

• NSW private health service providers: Adult health information is retained for at least 7 years after the last health service provided by the provider; information collected while the person was under 18 is retained until age 25. See IPC NSW guidance on section 25 of the HRIP Act.
• Victoria: Covered health service providers generally must not delete health information until more than 7 years after the last health service and, for information collected while the person was a child, after age 25, whichever is later. See Health Privacy Principle 4.2.
• ACT: Covered record keepers generally retain adult health information until more than 7 years after the last service, and information collected while the person was under 18 until after age 25. See Schedule 1, Principle 4.1.

Public-sector records, other jurisdictions and particular services may have different or longer requirements. Applicable legal holds and required disposal or transfer registers must also be preserved. A request to close an account does not override a clinical-record retention obligation.

9.2 Contact and business information

Contact information provided through partnership enquiries or access requests is retained for the duration of the business relationship and for a reasonable period afterward (up to 2 years), unless you request earlier deletion.

9.3 De-identified reasoning outputs

De-identified reasoning outputs (as described in Section 5.2) may be retained indefinitely, as they contain no personal information and serve the ongoing improvement of clinical and domain quality. These outputs are subject to periodic re-identification risk assessments.

9.4 Programme user data

Interaction data from programme users is retained for the period specified in the relevant Partner Organisation agreement, and is deleted or de-identified at the end of that period or on termination of the agreement, unless a longer period is required by law.

9.5 Direct MIA records and clinician account records

Direct MIA clinical records follow section 9.1 where the relevant health-records law applies. Clinician account information and security audit logs are retained only while needed for a lawful purpose, such as operating the account, security or accountability, or to satisfy a specific legal retention duty. They are not automatically assigned a clinical-record retention period. Where a log forms part of a clinical record, the applicable clinical-record requirement applies. Contact the Privacy Officer for access, correction, or information about retention and deletion.

9.6 Termination, deletion and backups

Account closure or contract termination triggers a review of whether personal information is still needed or legally required. Destruction or de-identification obligations apply even if no export is requested. Backup copies are included. Where immediate destruction is technically impossible, information must be put beyond use: it is not used or disclosed, others cannot access it, it remains secured, and reasonable steps are taken to destroy it when possible. Routine backup retention does not justify keeping information indefinitely. See OAIC guidance on APP 11.

12.1 Business and contact information

If you are a domain expert, researcher, or partner, you may contact us directly to request access (APP 12) or correction (APP 13) of your contact information. We will respond within 30 days.

12.2 Clinical information

For information you provide directly to UNCAPT through MIA, contact our Privacy Officer. For health information and reasoning outputs processed on behalf of a Partner Organisation, we recommend first contacting your treating clinician or healthcare provider. If you contact us directly, we will verify your identity and coordinate with the relevant Partner Organisation where applicable.

12.3 How to exercise your rights

To lodge a request for access or correction, please contact:

UNCAPT Privacy Officer
Email: info@uncapt.com
Level 4, 83 Mount St, North Sydney NSW 2060 Australia

If you are not satisfied with our response, you may lodge a complaint with the Office of the Australian Information Commissioner (OAIC).