AI Transparency Statement

Effective date: 14 October 2026

TrustShyft exists to keep a named person in the decision. Our customers use AI in their own work. TrustShyft records that a person with the right authority reviewed the AI’s output and made the decision, and keeps that record so it can be shown later.

What TrustShyft does

• Records who reviewed an AI-assisted output, what they decided, when, and the reasons they gave
• Keeps those records append-only, so entries cannot be edited or deleted through the platform
• Asks the reviewer for genuine engagement with the decision, not a one-click approval
• Lets customers search and report on their records for internal review, audit and regulator enquiries

What TrustShyft does not do

• Decide whether an AI output, a decision or an organisation is compliant
• Certify, approve or sign off on anything on a customer’s behalf
• Make decisions about any person’s employment, credit, legal matter or performance
• Replace the customer’s own judgement, policies or professional obligations

A TrustShyft record is evidence that a human made a decision. It is not a conclusion that the decision was right.

Data and models

The platform uses the OpenAI API, with processing in the United States, to help users review, summarise and report on their decision records. Records are sent to the model only to produce the output a user asked for. OpenAI does not use API inputs or outputs to train its models by default, and we have not opted in to data sharing. OpenAI may retain them for up to 30 days, for example to monitor for abuse. AI outputs are suggestions that a named user reviews before they are used.

Why the human step matters

AI can be wrong, and people reviewing AI can drift into approving it without really checking. That is a known behavioural effect, often called automation bias. TrustShyft is designed around it: the record captures that a person engaged with the decision, so the oversight is real, not a formality. Users remain responsible for the decisions they record.

For our customers’ own privacy disclosures

From 10 December 2026, organisations covered by the Privacy Act must describe in their privacy policies certain decisions that computer programs make, or substantially help make, using personal information. TrustShyft does not make those decisions, but the AI tools you use alongside it might. Your TrustShyft records can help you show where a person, rather than a program, made the decision. This summary may help with your own assessment:

• Personal information TrustShyft uses: your users’ identities and the decision details they record, which may name your clients or matters.
• What it produces: a record of human decisions and reports on those records, not decisions about individuals.
• Human involvement: every record is created by a named person.

Whether and how you need to disclose your AI use depends on how you use it, so take your own advice. Questions to privacy@trustshyft.com.au.