Frequently asked questions
This resource is designed to support clear decision‑making, promote transparency, and ensure that AI is used safely and ethically, whether you’re developing content, managing a project or procuring a service.
What are the benefits of labelling AI-generated content?
Show moreLabelling AI-generated content supports transparency in alignment with the WA Government AI Policy and the AI Toolkit for WA public sector staff. Even when labelling AI-generated content, agencies remain responsible for the accuracy, quality and appropriateness of published materials.
When to label AI-generated content
Agencies can encourage staff to label AI-generated content, particularly when it is made publicly available and may reasonably influence public understanding, perception, or trust in government information and services. This includes (but is not limited to): publications and reports, social media posts, public reports, websites, presentations, and images, audio or video content.
How to label AI-generated content
The purpose of labelling is to help audiences understand whether AI tools were used and the extent of human review or oversight applied to build public trust. Some examples are provided below:
- Disclaimer: AI was used in the preparation of this [document/ report/ webpage]. Full review and editorial control remain with [agency name].
- This [image/ video/ audio] is AI-generated. It has been reviewed by [agency name].
- This [image/ video/ audio] was generated using [AI tool name], using the prompt: ‘[prompt used]’. It has been reviewed by [agency name].
How can I mitigate AI risk when procuring a vendor?
Show moreThere are a number of model clauses available to support procurement of services where the vendor may be using AI in the provision of their services, or are developing an AI system.
- The Department of Treasury and Finance has included a clause on Generative Artificial Intelligence and Automated Decision Making in Clause 6.8 of Schedule 18 of the whole-of-government Request – ICT, CUA and/or Panel Arrangement template.
- The Commonwealth Government also has Artificial Intelligence model clauses, which can be adapted by agencies where appropriate.
Agencies should seek legal advice when drafting or negotiating AI-specific contract conditions.
Where possible, agencies are also encouraged to ensure that their procurements incorporate existing government policies and requirements, such as:
- The principles outlined in the WA Government AI Policy and relevant guidance and toolkits for the WA public sector. Note: Agencies remain responsible for completion of the WA Government Assurance Framework (Self-assessment) for relevant AI use cases.
- The WA Government Cyber Security Policy including the WA Government Offshoring Position.
Compliance with the Privacy Act 1988 (Cth) and (from 1 July 2026) the Privacy and Responsible Information Sharing Act 2024 when handling personal information.
How do I demonstrate return on investment for my AI project?
Show moreThe WA Government AI Policy requires that AI is used to benefit the community. It is recommended that agencies evaluate the benefits of their AI project through a return on investment evaluation prior to considering further investments to scale solutions.
- Outline the problem statement and solution for the AI
A clear problem statement establishes the baseline for assessing whether an AI solution is delivering its intended outcome and value. Define the problem, including key business and user needs, current pain points or inefficiencies, and their scale and impact across the organisation or service. Additionally, outline the role of the AI solution in addressing the identified problem.
Problem: A contact centre receives approximately 500 calls per day, operated by a limited number of staff during business hours only. Around half of enquiries are routine and repetitive, leading to capacity constraints, pressure on staff workload, and reduced access to timely information, particularly outside operating hours.
Solution: An AI‑enabled chatbot provides users with access to clear, curated answers to common questions sourced from approved documents, allowing routine enquiries to be resolved independently, including outside business hours.
- Articulate the intended benefits and value of the AI solution
Articulate the intended benefits and value of the AI solution addressing the problem. Benefits may include improved efficiency, service quality, wellbeing, access, equity, or environmental sustainability. Clearly defined benefits enable meaningful evaluation of whether the AI solution is delivering value relative to its cost and risk.
Intended benefits: The AI-enabled chatbot is expected to improve access to information through 24/7 availability, improve service quality and consistency, and reduce staff time spent responding to routine enquiries.
- Measure return on investment of the AI solution
Assess return on investment by comparing the costs of the AI solution with realised benefits over time, using baseline metrics and regular performance reviews against expected financial and non-financial outcomes.
Cost assessments should encompass technical and operational expenses, as well as ongoing human effort for validation, quality assurance, oversight and governance. Broader benefits beyond direct efficiency gains should also be considered, including opportunity cost, workload redistribution, staff time savings and service quality.
The Department of Treasury and Finance has developed the Total Cost of Ownership and Benefits Model as part of their Strategic Asset Management Framework, which is a useful resource for outlining cost and benefits of an AI project. pressure and demonstrated return on investment.
Return on investment: The AI-enabled chatbot resolved approximately 200 routine queries per day and provided users with 24/7 access to accurate information outside business hours. Staff surveys indicated that 70% of call centre staff spent less time responding to repetitive enquiries and were able to focus more on complex questions. These outcomes eased staff pressure and demonstrated return on investment.
I am an AI Accountable Officer, what do I do?
Show moreThis guidance outlines the responsibilities and key actions expected of an AI Accountable Officer.
Under the WA Government AI Policy and AI governance toolkit for WA public sector entities, an AI Accountable Officer:
- Ensures their entity’s AI usage aligns with the WA Government AI Policy and existing information governance obligations including: the State Records Act 2000 (WA), Privacy and Responsible Information Sharing Act 2024 (WA), Public Sector Management Act 1994 (WA), Freedom of information Act 1992 (WA), and WA Information Classification Policy.
- Approves AI use case self-assessments as they are submitted through the WA Government AI Assurance Framework.
- Ensures that the risks of using AI systems and tools within the entity are effectively managed.
- Supports staff to understand the opportunities and risks of AI tools.
- Is the contact point for the Office of Digital Government (DGov), and shares information with DGov on possible risks and opportunities in AI use in their entity.
- Ensures oversight of all AI projects and tools in use in the entity through the AI use case register and AI tools reference list.
- Keeps up to date with governance requirements.
The WA Government AI Assurance Framework sets out the following self-assessment and approval requirements for the AI Accountable Officer:
- A self-assessment must be completed for non-trivial AI and automated decision-making systems across their life cycle, including low risk use cases.
- Upon submission, the self-assessment is shared with the project team lead, the AI Accountable Officer, and the Office of Digital Government.
- The AI Accountable Officer must review and either approve or reject the project within three weeks.
When reviewing a self-assessment, the AI Accountable Officer should:
- Confirm the self-assessment is complete and clearly articulated.
- Verify the project phase and ensure re-assessments are submitted before each subsequent phase. Self-assessments can be updated using the link emailed to the project’s primary contact.
- Ensure data use, storage, and processing arrangements comply with WA’s Data Offshoring Position, Cyber Security Policy and Privacy and Responsible Information Sharing Act.
- Ensure appropriate risk controls are in place, including human validation and oversight, incident management, monitoring and performance reviews, and clear accountability for operations and change management
- Verify that risks have been appropriately assessed, mitigated and clearly justified.
Print this page
