APS AI Tracker

OCO

Office of the Commonwealth Ombudsman

View the original ↗

Tracked since 12 Nov 2025 · 2 changes of substance across 3 captures · last changed 4 May 2026

The story so far

How to read this

  1. Tracking begins. The statement itself says it was first published 27 Feb 2025.

  2. substantive Discloses two new AI uses --- coding assistants and Azure text-to-speech for training materials --- alongside existing security monitoring, and logs the December 2025 update.

    read by Claude Opus 5
    • New disclosure: AI coding assistants for development, configuration and testing, limited to digital specialists with mandatory testing before implementation
    • New disclosure: Azure AI Text to Speech used to produce internal training materials from human-written scripts
    • States no direct public access to the security monitoring AI or coding-assistant-supported systems
    • Changelog entry added for 20 December 2025; initial release date corrected to 27 February 2025
    • Link to the Automated Decision-making Better Practice Guide points to a different (older) PDF
    • New commitment (human oversight): All human and AI-assisted coding must be tested for accuracy before implementation
    • Safeguard added: testing or evaluation
    • Public-facing AI use: not addressed → none
    • Stated first-published date: 2025-02-26 → 2025-02-27
    • Stated last-updated date: — → 2025-12-20
    Show the words that changed ↓

  3. substantive Rewrites the statement for policy v2.0, disclosing AI complaint summarisation and form population, adding prohibited uses and a Chief AI Officer, and dropping the annual review commitment.

    read by Claude Opus 5
    • New disclosure: AI summarises complaints for allocation and triage (not for deciding action)
    • New disclosure: AI populates complaint forms from letters/emails, confirmed by an officer
    • Adds prohibited uses: no AI statutory decisions, no AI-generated external communications
    • Chief AI Officer named as approver for low/medium risk use cases; Executive Committee for high risk (none currently)
    • Removed: commitment to review the statement annually or when use cases change
    • Removed: detailed list of what future use-case disclosures would cover (legal authority, remediation, human rights, privacy)
    • Removed: reference to the Automated Decision-making Better Practice Guide and the Microsoft Copilot/Azure pilots
    • Maps use to workplace productivity pattern, service delivery and corporate/enabling domains
    • Contact line now invites complaints about the Office's AI use
    • Links updated to policy and Standard version 2.0
    • Commitment dropped (human oversight): All human and AI-assisted coding must be tested for accuracy before implementation
    • Commitment dropped (human oversight): Each AI use case requires Ombudsman's approval after ITGC endorsement
    • Named tool dropped: Microsoft Azure AI
    • Named tool dropped: Microsoft Copilot
    • Commitment dropped (will): Statement reviewed annually or when AI use cases and usage change
    • Commitment dropped (will): Will be transparent in assessment, preparation, engagement, adoption and monitoring of AI
    • Safeguard dropped: testing or evaluation
    • Chief AI Officer: not mentioned → in place
    • Domain added: service delivery
    • New commitment (human oversight): AI-populated complaint form suggestions must be confirmed by an Officer before being recorded
    • New commitment (human oversight): Each use case requires individual approval after impact assessment; high risk goes to Executive Committee
    • New commitment (human oversight): Statutory decisions are always made by an Officer with the appropriate delegation
    • New commitment (will not): Complaint summaries are never used to make a decision about action on a complaint
    • New commitment (will not): Staff must not use AI to generate text for communications outside the Office
    • Safeguard added: human review of outputs
    • Usage pattern added: workplace productivity
    • Policy version referenced: v1 → v2
    • Public-facing AI use: none → with human review
    • Review cadence: annual and on change → on change
    • Stated last-updated date: 2025-12-20 → 2026-04
    Show the words that changed ↓

What the statement says

How to read this

The Office of the Commonwealth Ombudsman says it uses AI to summarise complaints for allocation and triage, populate complaint forms from correspondence (confirmed by an Officer), monitor ICT security, assist code completion, and generate text-to-speech training materials. An internal AI policy requires individual approval of each use case after impact assessment (Chief AI Officer for low/medium risk, Executive Committee for high risk), and AI is never used to make statutory decisions or to write communications to complainants. read by Claude Opus 5

What it says AI is used for

  • Workplace productivity
  • Service delivery
  • Corporate and enabling

Public-facing AI: yes, with human review

Safeguards named: risk assessment, human review of outputs, staff training, privacy or security controls, a governance body, an acceptable-use policy

Against the Standard 6/8

  • Intentions behind AI use
  • Use classified by DTA usage pattern or domain
  • Public-facing use addressed
  • Monitoring and protection measures
  • Compliance with the policy
  • Compliance with legislation
  • Date last updated
  • Public contact

What the policy requiresPolicy v2.0

Accountable official
Chief Operating Officer and Chief Information Officer
Strategic position on AI
not mentioned
AI use-case register
not mentioned
Staff training
available
Policy version referenced
version 2.0

What the AI Plan asksAI Plan

Chief AI Officer
in place (Chief AI Officer)

Currency as of 29 Aug 2026

Review cadence
when the approach changes
Says it was last updated
Apr 2026
Last change we observed
4 May 2026
Updated since policy 2.0
yes
Annual review
within a year of its own date

Commitments

  • will notDoes not currently use or intend to use AI to make discretionary decisions
  • will notComplaint summaries are never used to make a decision about action on a complaint
  • will notStaff must not use AI to generate text for communications outside the Office
  • human oversightAI-populated complaint form suggestions must be confirmed by an Officer before being recorded
  • human oversightStatutory decisions are always made by an Officer with the appropriate delegation
  • human oversightEach use case requires individual approval after impact assessment; high risk goes to Executive Committee
  • willWill update this statement to outline any additional AI use cases and technologies implemented

The statement

How to read this

The Digital Transformation Agency (DTA)'s Policy for the responsible use of AI in government sets out the Australian Government approach to Artificial Intelligence (AI).

The Office of Commonwealth Ombudsman (the Office), including the ACT Ombudsman, is committed to identifying opportunities for AI to help us deliver our mission of “helping people, improving government”.

Any use of AI by the Office is carefully considered for compatibility with Australia’s AI Ethics Principles and the Policy for the Responsible Use of AI in Government. These require transparency in how AI is used by the Office.

  • the scope and classification system for AI use in the Office
  • ways we do and don’t use AI in the Office
  • the policy and governance for AI use in the Office.

The Office currently uses specific AI tools to make our work more efficient. This helps staff to focus on more complex and meaningful work. Currently, the Office uses AI for the following activities:

  • To create a short summary of complaints to the Office. This summary is used to help us find the right team and staff member to assist with the complaint (“allocation”), and to help us prioritise the most urgent complaints (“triage”). These summaries are never used to make a decision about what action to take in relation to a complaint.
  • To populate complaint forms from information we receive by letter or email. This helps minimise manual data entry. These suggestions must be confirmed by an Officer before they are recorded.
  • To support security monitoring of our ICT systems. This AI is embedded into the monitoring software we use, and uses machine learning to help us identify new security threats.
  • To support development and programming. This is limited to staff involved in developing new ICT capabilities in the Office and supports them with code completion.
  • We use Text-to-Speech software to support the creation of internal training materials based on a human generated text script.

If we decide to use AI in new ways in the future, we will update this statement to explain how this may affect you.

The Digital Transformation Agency publishes a classification system for the use of AI in government. In this system, the Office’s use of AI primarily falls into the workplace productivity pattern. This means we may use AI to automate routine tasks, manage workflows and to support triage and categorisation of communications.

The Office uses AI in the service delivery domain, to support staff to register, categorise and triage complaints. We do not use AI in making decisions. We also use AI in the corporate and enabling domain, including HR, finance, media, communications, and IT.

Certain uses of AI are prohibited within the Office. We do not allow staff to use AI tools to make decisions under our legislation. Statutory decisions are always made by an Officer with the appropriate delegation.

We also do not allow staff to use AI to generate text to be used in communications outside the Office. Our communications with complainants about their complaint are always written by an Officer.

AI usage policy and processes

We have an internal AI Policy that aligns with advice and guidance provided by the DTA and other agencies for using AI services responsibly.

  • we have a framework to adopt AI where it can support us to do our work more effectively and efficiently, consistent with Australia’s AI Ethics Principles
  • our AI use has appropriate governance
  • our staff are appropriately trained in AI systems
  • relevant risks are identified and appropriately treated
  • our use of AI supports our reputation and does not compromise stakeholder trust in our work.

Each use case requires individual approval after a detailed impact assessment. Low and medium risk use cases can be approved by the Chief AI Officer. High risk use cases are considered by the Office's Executive Committee. The Office does not currently engage in any high-risk use of AI.

The Chief Operating Officer and Chief Information Officer are the designated AI Accountable Officials.

For questions about this statement or if you would like to make a complaint about how the Office has used AI you can do so by writing to: ai@ombudsman.gov.au.

Date| Note
---|---
February 2025| Initial release.
December 2025| Add AI Assistant for Code Completion and Programming Usage and Text to Speech for training materials.
April 2026| Update to reflect revised AI Policy and new use cases.

[1] Explanatory memorandum on the updated OECD definition of an AI system

Statement text © Office of the Commonwealth Ombudsman, reproduced for transparency tracking (most agency content is CC BY 4.0 — check the original for specifics).

Every revision

Every capture the daily scrape recorded that differed from the last, including edits with no change of substance. Pick a dot to view the statement as it stood at that capture.

  1. substantive-509

    Rewrites the statement for policy v2.0, disclosing AI complaint summarisation and form population, adding prohibited uses and a Chief AI Officer, and dropping the annual review commitment.

    read by Claude Opus 5
    View diff
    # AI Transparency Statement ### Introduction The Digital Transformation Agency (DTA)'s [ _Policy for the responsible use of AI in government_](https://www.digital.gov.au/sites/default/files/documents/2024-085-12/Policy%20for%20the%20responsible%20use%20of%20AI%20in%20gGovernment%20v1.12.0_0.pdf) sets out the Australian Government approach to embrace the opportunities of Artificial Intelligence (AI) 1 and provide for safe and responsible use of AIArtificial Intelligence (AI). The Office of Commonwealth Ombudsman (the Office), including the ACT Ombudsman, is committed to identifying ethical, responsible and meaningful usage of AI toopportunities for AI to help us deliver our mission of _“helping people, improving government”_. We willAny use of AI bey transparent in our assessment, preparation, engagement, adoption, monitoring, and pivot to changes in AI technology, environment and policy requirements. ### Scope and Usage Currently, the Office uses AI for the following activities: 1. Support security monitoring of its ICT systems. This AI is embeddedhe Office is carefully considered for compatibility with Australia’s AI Ethics Principles and the Policy for the Responsible Use of AI in Government. These require transparency in how AI is used by the Office. This statement explains: - the scope and classification system for AI use into the monitoring software we use, with access limited to ICT and Cyber Security personnel that are suitably cleared, trained, and have a need to know and need to access the monitoring system. No direct public access. 1. AI Assistants to support code completion and programming. This AI is used to assist with developing, configuring, maintaining and testing commercial softwareOffice - ways we do and don’t use AI in the Office - the policy and governance for AI use in the Office. ### Tools in use The Office currently uses specific AI tools to make our work more efficient. This helps staff to focus on more complex and meaningful work. Currently, the Office uses. Access is limited to digital specialists with suitable ICT and Data skills. The use of these AI Coding Assistants is only for approved use cases of the Office’s system. All Human and AI assisted coding and system development, must undergo testing to ensure accuracy before being implement as part of the Office technical change management processes. The public does noI for the following activities: 1. To create a short summary of complaints to the Office. This summary is used to help us find the right team and staff member to assist with the complaint (“allocation”), and to help us prioritise the most currgently access systems that are supported by the complaints (“triage”). These summaries are never used tools. 1. Use of Azure AI Text to Speech to support the c make a decision about what action to take in relation of internal training materials based on a precise human generated text script. Access is limited to digital specialists with suitable ICT and Data skills. The Office has also completed pilots of two Microsoft AI capabilities – Microsoft Co-Pilot and Microsoft Azure AI services. The Azure AI pilot, within the Office secure IT environment, tested the feasibility of using Azure AI services to summarise written complaints for decision makers to review in order to process them. If we do implement any additional AI use cases and technologies, we will updato a complaint. 1. To populate complaint forms from information we receive by letter or email. This helps minimise manual data entry. These suggestions must be confirmed by an Officer before they are recorded. 1. To support security monitoring of our ICT systems. This AI is embedded into the monitoring software we use, and uses machine learning to help us identify new security threats. 1. To support development and programming. This is limited thiso statement to outline our use, with a summary of: - why the Office is using AI - the legislative authority for the use of AI - whether the public may directly interact with the AI or be significantly impacted by it - how the Office intends to notify those affected by our use of AI - what role AI plays in relation to decision-making, administrative action or service delivery, or other such usage patterns and domains as described in tff involved in developing new ICT capabilities in the Office and supports them with code completion. 1. We use Text-to-Speech software to support the creation of internal training materials based on a human generated text script. If we decide to use AI in new ways in the future, we will update this statement to explain how this may affect you. ### Scope and classification The Digital Transformation Agency publishes [Ca classifications system for AI use](https://www.digital.gov.au/policy/ai/resources/use-classification) - measures to identify and protect the public against negative impacts and other risk mitigation measures - measures in place to identify and remediate errors - how the use complies with administrative law principles, is consistent with human rights obligations, applicable legsites/default/files/documents/2025-12/Standard%20for%20AI%20transparency%20statements%202.0_0.pdf) for the use of AI in government. In this system, the Office’s use of AI primarily falls into the **workplace productivity** pattern. This means we may use AI to automate routine tasks, manage workflows and to support triage and categorislation including the [Privacy Act 1988](https://www.legislation.gov.au/C2004A03712/latest/text) and the [Protective Security Policy Framework](https://www.protectivesecurity.gov.au) - compliance with each requirement under the [Policy for the responsible use of AI in government](https://www.digital.gov.au/sites/default/files/documents/2024-08/Policy%20for%20the%20responsible%20use%20of%20AI%20in%20government%20v1.1.pdf) _._ Tof communications. The Office uses AI in the **service delivery** domain, to support staff to register, categorise and triage complaints. We do not use AI in making decisions. We also use AI in the **corporate and enabling** domain, including HR, finance, media, communications, and IT. ### Prohibited use Certain uses of AI are prohibited within the Office. We does not currently use or intendallow staff to use AI tools to make discretionary decisions in accordance with the best practices set out in the Office’s [Automated Decision-making Better Practice Guide](https://www.ombudsman.gov.au/__data/assets/pdf_file/0029/288236/OMB1188-Automated-Decision-Making-Report_Final-A1898885.pdf). ### Governance **AI Governance**\ Each AI use case and AI technology in the Office requires the Ombudsman’s approval, after endorsement by the Information Technology Governance Committee (ITGC) chaired by the Deputy Ombudsman. ecisions under our legislation. Statutory decisions are always made by an Officer with the appropriate delegation. We also do not allow staff to use AI to generate text to be used in communications outside the Office. Our communications with complainants about their complaint are always written by an Officer. ### **AI Uusage Ppolicy and Pprocesses**\ We have an internal AI Policy that aligns with advice and guidance provided by the DTA and other agencies for using AI services responsibly. We have processes to ensure: - our AI use is appropriately governed - our engagement with AI is confident, safe and responsible - our staff are appropriately trained in AI - any relevant and likely cyber, data and privacy risks are identified and addressed - our AI access and usage is monitored - our stakeholders have trust in our use of AI. ______________________________________________________________________ **This policy helps ensure: - we have a framework to adopt AI where it can support us to do our work more effectively and efficiently, consistent with Australia’s AI Ethics Principles - our AI use has appropriate governance - our staff are appropriately trained in AI systems - relevant risks are identified and appropriately treated - our use of AI supports our reputation and does not compromise stakeholder trust in our work. Each use case requires individual approval after a detailed impact assessment. Low and medium risk use cases can be approved by the Chief AI Officer. High risk use cases are considered by the Office's Executive Committee. The Office does not currently engage in any high-risk use of AI. ### Accountable Officials**\ The Chief Operating Officer and Chief Information Officer are the designated AI Accountable Officials. **Transparency Statement Updates**\ This statement will be reviewed annually or when we change our AI use cases and usage. ### Contact For questions about this statement or for further information onif you would like to make a complaint about how the Office’ has usage of AI, please contacted AI you can do so by writing to: [ai@ombudsman.gov.au](mailto:ai@ombudsman.gov.au). ### Change Llog ______________________________________________________________________ **Date**| **Note**\ ---|---\ 27 February 2025| Initial release.\ 20 December 2025| Add AI Assistant for Code Completion and Programming Usage and Text to Speech for training materials.\ April 2026| Update to reflect revised AI Policy and new use cases. [1] [Explanatory memorandum on the updated OECD definition of an AI system](https://www.oecd.org/en/publications/explanatory-memorandum-on-the-updated-oecd-definition-of-an-ai-system_623da898-en.html) Further information: [ Complaints ](https://www.ombudsman.gov.au/complaints) [ Home ](https://www.ombudsman.gov.au/home)
    a6ee3fc
  2. substantive+918

    Discloses two new AI uses --- coding assistants and Azure text-to-speech for training materials --- alongside existing security monitoring, and logs the December 2025 update.

    read by Claude Opus 5
    View diff
    # AI Transparency Statement ### Introduction The Digital Transformation Agency's [ _Policy for the responsible use of AI in government_](https://www.digital.gov.au/sites/default/files/documents/2024-08/Policy%20for%20the%20responsible%20use%20of%20AI%20in%20government%20v1.1.pdf) sets out the Australian Government approach to embrace the opportunities of Artificial Intelligence (AI) 1 and provide for safe and responsible use of AI. The Office of Commonwealth Ombudsman (the Office), including the ACT Ombudsman, is committed to identifying ethical, responsible and meaningful usage of AI to deliver our mission of _“helping people, improving government”_. We will be transparent in our assessment, preparation, engagement, adoption, monitoring, and pivot to changes in AI technology, environment and policy requirements. ### Scope and Usage Currently, the Office uses AI to sfor the following activities: 1. Support security monitoring of its ICT systems. This AI is embedded into the monitoring software we use, with access limited to ICT and Cyber Security personnel that are suitably cleared, trained, and have a need to know and need to access the monitoring system. The Office has recently No direct public access. 1. AI Assistants to support code completion and programming. This AI is used to assist with developing, configuring, maintaining and testing commercial software the Office uses. Access is limited to digital specialists with suitable ICT and Data skills. The use of these AI Coding Assistants is only for approved use cases of the Office’s system. All Human and AI assisted coding and system development, must undergo testing to ensure accuracy before being implement as part of the Office technical change management processes. The public does not currently access systems that are supported by these tools. 1. Use of Azure AI Text to Speech to support the creation of internal training materials based on a precise human generated text script. Access is limited to digital specialists with suitable ICT and Data skills. The Office has also completed pilots of two Microsoft AI capabilities – Microsoft Co-Pilot and Microsoft Azure AI services. The Azure AI pilot, within the Office secure IT environment, tested the feasibility of using Azure AI services to summarise written complaints for decision makers to review in order to process them. If we do implement any additional AI use cases and technologies, we will update this statement to outline our use, with a summary of: - why the Office is using AI - the legislative authority for the use of AI - whether the public may directly interact with the AI or be significantly impacted by it - how the Office intends to notify those affected by our use of AI - what role AI plays in relation to decision-making, administrative action or service delivery, or other such usage patterns and domains as described in the [Classifications system for AI use](https://www.digital.gov.au/policy/ai/resources/use-classification) - measures to identify and protect the public against negative impacts and other risk mitigation measures - measures in place to identify and remediate errors - how the use complies with administrative law principles, is consistent with human rights obligations, applicable legislation including the [Privacy Act 1988](https://www.legislation.gov.au/C2004A03712/latest/text) and the [Protective Security Policy Framework](https://www.protectivesecurity.gov.au) - compliance with each requirement under the [Policy for the responsible use of AI in government](https://www.digital.gov.au/sites/default/files/documents/2024-08/Policy%20for%20the%20responsible%20use%20of%20AI%20in%20government%20v1.1.pdf) _._ The Office does not currently use or intend to use AI to make discretionary decisions in accordance with the best practices set out in the Office’s [Automated Decision-making Better Practice Guide](https://www.ombudsman.gov.au/__data/assets/pdf_file/0025/317437/9/288236/OMB1188-Automated-Decision-Making-Better-Practice-Guide-March-202Report_Final-A1898885.pdf). ### Governance **AI Governance**\ Each AI use case and AI technology in the Office requires the Ombudsman’s approval, after endorsement by the Information Technology Governance Committee (ITGC) chaired by the Deputy Ombudsman. **AI Usage Policy and Processes**\ We have an internal AI Policy that aligns with advice and guidance provided by the DTA and other agencies for using AI services responsibly. We have processes to ensure: - our AI use is appropriately governed - our engagement with AI is confident, safe and responsible - our staff are appropriately trained in AI - any relevant and likely cyber, data and privacy risks are identified and addressed - our AI access and usage is monitored - our stakeholders have trust in our use of AI. ______________________________________________________________________ **Accountable Officials**\ The Chief Operating Officer and Chief Information Officer are the designated AI Accountable Officials. ______________________________________________________________________ **Transparency Statement Updates**\ This statement will be reviewed annually or when we change our AI use cases and usage. ### Contact For questions about this statement or for further information on the Office’s usage of AI, please contact [ai@ombudsman.gov.au](mailto:ai@ombudsman.gov.au). ### Change Log ______________________________________________________________________ **Date**| **Note**\ ---|---\ 267 February 2025| Initial release.\ 20 December 2025| Add AI Assistant for Code Completion and Programming Usage and Text to Speech for training materials. [1] [Explanatory memorandum on the updated OECD definition of an AI system](https://www.oecd.org/en/publications/explanatory-memorandum-on-the-updated-oecd-definition-of-an-ai-system_623da898-en.html) Further information: [ Complaints ](https://www.ombudsman.gov.au/complaints) [ Home ](https://www.ombudsman.gov.au/home)
    078058e
  3. first tracked+4938

    First tracked revision.

    9226d49