28 February 2025

Introduction

The Parliamentary Budget Office (PBO) is committed to the responsible and ethical use of Artificial Intelligence (AI) technologies.

This AI transparency statement outlines the PBO's approach to AI, aligned with the broader Commonwealth direction and the Parliament of Australia Digital Strategy 2023-2027.

This statement also includes guidance on the role the Department of Parliamentary Services (DPS) plays as the ICT shared services provider for the Parliamentary departments and the Parliamentary Workplace Support Service (PWSS), and the shared approach to the responsible use of AI and compliance with the Digital Transformation Agency (DTA) Policy for the responsible use of AI in government.

Accountable Official

The PBO’s Chief Operating Officer (COO) is the PBO’s accountable official, appointed in line with the DTA standard for accountable officials.

Review and Updates of this AI Transparency statement

This transparency statement will be reviewed annually or when significant changes to the use or governance of AI use occur.

This statement was last updated on 28 February 2025.

AI Adoption and Use

The PBO believe that as with any technology, AI presents opportunities that should be harnessed and risks that must be managed.

The PBO has been proactive in its approach to the responsible, safe and ethical use of AI and governance of that use. We have implemented arrangements for managing the use of AI in line with the following principles:

  • Leveraging the benefits of AI in a safe and secure way. This means having robust governance processes on the ethical adoption of AI tools to ensure benefits are achieved and risks are managed, while maintaining confidentiality of information, data security, and that systems perform as expected.
  • Transparency and accountability in our use of AI. PBO staff must be able to explain, justify and take ownership of advice and decisions when utilising AI.
  • Clarity in roles and responsibilities. Having clear accountabilities for the adoption of AI and understanding its use.
  • Sustainability in approach. We will invest in building AI capability for the long term, to capture the business benefits from the capability in a safe way.

Integration of AI into the operations of the PBO is guided by:

The PBO’s approach to the adoption of AI is designed to:

  • harness business benefits whilst effectively mitigating and managing risks align with the Parliament of Australia's Digital Strategy
  • meet PBO client and staff expectations for modern ICT experiences
  • reduce the risk associated with the use of AI
  • support innovation and unlocking new value using AI, and
  • support whole-of-government direction and policy on the use of AI.

The PBO collaborates closely with the other Parliamentary departments (Senate, House of Representatives and DPS) and Parliamentary Workplace Support Service (PWSS) in a shared responsibility model for the use of AI.

How the PBO uses AI

The PBO follow the Organisation for Economic Cooperation and Development’s (OECD) definition of what constitutes AI, and considers the risks and benefits of its use against the DTA classification system for AI.

Current use of AI at the PBO does not directly impact our clients or the Australian public. It is currently limited to usage that enhances workplace productivity in the domains of corporate and enabling and supporting service delivery. Common uses include summarising meetings, informing the development of documents or code, automating some routine tasks and improving the ability to search for information.

Going forward, the PBO will further consider and review AI tools for use cases that enhance our work and create value for our clients and stakeholders with a strong focus on security and confidentiality of information.

The PBO is committed to being transparent about its use of AI. While the PBO uses AI internally and does not currently use AI to communicate directly with clients or stakeholders, the PBO also commits to always ensuring it is obvious to individuals if:

  • they are interacting with an AI system, such as an AI chatbot, or
  • the content is completely AI produced, without human review.

The PBO outsources ICT service provision to DPS. As the ICT service provider to the Parliament, the Parliamentary departments and Parliamentary Workplace Support Service (PWSS), DPS provides the overarching environment and framework for AI. The DPS AI Transparency statement includes information on how DPS uses AI to provide ICT services to the PBO.

Parliamentary Shared Responsibility Model

The PBO, in partnership with DPS will meet the requirements under the DTA responsible use of AI for Government policy through a shared responsibility model.

The shared responsibility model recognises:

  • each Parliamentary department and PWSS is independent, but share a common service provider, DPS, that provides the core ICT services for the Parliament, and
  • the use of AI by an individual or group within the parliament may impact other individuals, groups, organisations or the parliament as a whole; therefore collaboration is essential.

The shared responsibility model is managed by an AI working group composed of one or more representatives from each organisation, and coordinated by DPS ICT.

See the DPS AI Transparency statement for more information on how DPS support parliament’s use of AI and the implementation of the shared responsibility model.

AI Safety and Monitoring

The PBO have implemented controls to manage the risks associated with AI tools, and generative AI in particular, including:

  • Ensuring staff complete general AI training, and that training for specific roles or uses can be made available
  • Policies and guidance to staff to never upload or input any sensitive or classified information into public AI platforms
  • Preventing access to non-approved AI platforms
  • Displaying reminders to staff before accessing approved platforms of organisational expectations on their use
  • Requiring that all AI use is formally approved and registers maintained of:
    • staff that have been approved to use particular AI tools and,
    • AI tools that have been approved for broad use and particular use cases.

DPS plays a central role in ensuring the responsible and ethical use of AI technologies within the Parliament.

DPS employs the Commonwealth AI assurance framework to support the safe and ethical use of AI. DPS monitors, logs and reports on the use of:

  • DPS managed AI solutions, and
  • third-party AI used by the Parliament.

When the PBO wishes to adopt and use a new AI tool it is required that:

  • new AI use cases or requests to use AI are reported to the PBO ICT Committee
  • the PBO COO or a delegate will inform the Parliamentary AI Working Group and DPS ICT as appropriate before using AI
  • that risks are identified and the need for mitigations considered
  • testing is undertaken to evaluate whether AI is suitable for the intended use case
  • the value and effectiveness of adopted AI solutions is monitored and assessed against their use case, and adjusted as necessary
  • legal and regulatory compliance is maintained

PBO staff that use AI tools must:

  • report unintended outcomes, faults or negative impacts from AI use, and
  • report if sensitive information is inadvertently entered into Generative AI tools or services.

All unintended outcomes or faults with AI systems used by the PBO should be reported to the PBO COO. They will also be raised with DPS IT and DPS Cyber Security via internal governance and processes, where appropriate. Staff are also encouraged to raise issues directly with DPS IT and DPS Cyber Security.

Contact

For further information about the adoption of artificial intelligence within the Parliamentary Budget Office, please contact us.