AI Terms

Version: 1.0 

Effective Date: 12 August 2026

Previous versions : N/A

AI.1 Definitions

“Order Document” means the project schedule, statement of work or order under which the Services are supplied and which incorporates these AI provisions (the “AI Terms”), whether by setting them out as an annex or by incorporating them by reference.

“AI System” means any artificial intelligence, machine learning, generative or agentic system used by the Supplier in providing the Services under the Order Document.

“AI Output” means any content, recommendation, classification or result generated by an AI System.

“Automated Decision” means a decision produced by an AI System without human review.

“Human Oversight” means review of AI Outputs or Automated Decisions by a competent person, at the level applicable to the activity under the Supplier’s AI Human Oversight Matrix published at www.cubesys.com.au/AIHumanOversight (the “Oversight Matrix”), before they are relied upon or implemented. These definitions build on “Services”, “Deliverables”, “Client Data” and “Protected Information” as defined in the Agreement.

AI.2 Transparency and disclosure of AI use

The Supplier will inform the Client where AI Systems are used in delivering the Services, their intended purpose, the categories of Client Data they process, and any material limitations, and will maintain and make available to the Client on reasonable request a register of the AI Systems used in the engagement. On Handover of an AI System within the meaning of clause AI.12, the Supplier will provide the Client with an extract of the register for that AI System as at the date of Handover.

AI.3 Responsibility for AI Outputs

AI Outputs are provided to support, not replace, the Client’s judgement. The Client remains responsible for decisions it makes in reliance on AI Outputs following its own review. The Supplier is responsible for configuring, operating and governing the AI Systems it operates in providing the Services with reasonable care and skill and in accordance with good industry practice, and will not implement an Automated Decision that materially affects the Client or an individual without the Human Oversight applicable to that activity under the Oversight Matrix. The Supplier will not implement an Automated Decision that produces legal effects concerning an individual, or that otherwise significantly affects an individual, without Human Oversight by a competent person.

AI.4 Human oversight and guardrails

For agentic workflows, the Supplier will maintain human-in-the-loop checkpoints, guardrails and testing before an agent is released into the Client’s environment, and will operate defined controls over what each agent can access and do. The Supplier will apply the oversight approach set out in the Oversight Matrix, as updated from time to time to reflect good industry practice, provided that no update will materially reduce the protections given to the Client under these AI Terms.

AI.5 AI data use and grounding

The Supplier will use Client Data and Protected Information only to ground and operate the AI Systems for the purpose of providing the Services to the Client. The Supplier will not use Client Data or Protected Information to train foundation models, or for the benefit of any other client, and will not transfer it outside Australia without the Client’s prior written approval, other than processing inherent in the underlying cloud platforms of the disclosed AI Systems as described in the register maintained under clause AI.2.

AI.6 Responsible and intended use

The Supplier will use AI Systems only for their approved intended purpose within the Services, in a lawful and responsible manner consistent with the Supplier’s AI policy (cubesys-pol-18) and applicable law.

AI.7 AI impact and risk assessment

Before deploying an AI System that processes the Client’s sensitive information or that may significantly affect the Client or an individual, and on any material change, the Supplier will assess the potential impacts of that AI System on the Client and on affected individuals, and will make a summary of the assessment available to the Client on reasonable request.

AI.8 AI incident notification

The Supplier will notify the Client promptly after becoming aware of any AI incident that materially affects the Client’s data, the Services, or the reliability or safety of AI Outputs relied upon by the Client, and will cooperate to mitigate it. This clause operates together with the Supplier’s incident management process (cubesys-pol-14) and the privacy and confidentiality obligations in the Agreement.

AI.9 Third-party AI providers

Where AI capability is provided by third parties (including Microsoft), the Supplier remains responsible for the Services it is engaged to provide and will use reasonable endeavours to engage such providers on terms consistent with these AI Terms. The Supplier is not responsible for the acts, omissions or terms of a third-party provider to the extent they are outside the Supplier’s reasonable control, provided the Supplier has complied with this clause.

AI.10 Assurance and audit

The Supplier operates management systems certified to ISO 9001 and ISO 27001 and an AI management system aligned with ISO/IEC 42001, and will make evidence of its current certification status available to the Client on reasonable request. A change in certification status does not of itself constitute a breach of these AI Terms, provided the Supplier continues to operate materially equivalent controls. The Client’s audit rights under the Agreement extend to the Supplier’s use of AI Systems in providing the Services, subject to the conditions of the audit clause in the Agreement.

AI.11 Application and precedence

These AI Terms apply where the Supplier uses an AI System in delivering the Services under the Order Document, and apply only to the Services provided under that Order Document. Where the Order Document incorporates these AI Terms by reference to a published version, the version in force at the date of the Order Document applies for the term of that Order Document, and the Supplier will retain and make available on reasonable request each superseded version. These AI Terms are incorporated into the Agreement through the Order Document and do not otherwise amend the executed terms of the Agreement. Where these AI Terms conflict with the Agreement, the Agreement prevails in accordance with clause 1.4 of the Agreement, except to the extent these AI Terms expressly govern the use of AI Systems in the Services, in which case these AI Terms prevail for that subject matter only. Clauses AI.5 and AI.8 survive termination or expiry of the Order Document in respect of Client Data processed by an AI System during the term. Clause AI.12 applies on and from Handover and survives termination or expiry of the Order Document.

AI.12 Handover and end of term

Where an AI System configured or deployed by the Supplier remains in the Client’s environment after Handover, or after expiry or termination of the Order Document, responsibility for the ongoing operation, monitoring, maintenance and governance of that AI System rests with the Client from that time, unless the Supplier is separately contracted to provide those services. “Handover” means the date on which the Supplier makes an AI System available to the Client to operate or manage, and notifies the Client that it has done so. On or about Handover the Supplier will provide the Client with a handover record identifying the AI System, its configuration, its intended purpose, its known limitations, the tier applicable to its activities under the Oversight Matrix, and the extract of the register for it under clause AI.2; a delay or deficiency in that record does not postpone Handover. From Handover the Supplier’s obligations under clauses AI.4, AI.6 and AI.7 cease in respect of that AI System, and the Supplier is not responsible for the behaviour, outputs or performance of that AI System to the extent arising from events occurring after Handover, including any change made to the AI System, its configuration, its permissions or its data sources by or on behalf of the Client, any change to an underlying platform, or use of the AI System for a purpose other than the intended purpose recorded at Handover. Nothing in this clause limits the Supplier’s responsibility for its acts or omissions before Handover, or affects any right or remedy of the Client under the Australian Consumer Law or other law which cannot lawfully be excluded, restricted or modified.