IP Australia

1 of 6

IP Australia artificial intelligence and automated decision-making transparency statement This transparency statement provides the public with relevant information about IP Australia’s use of artificial intelligence (AI) and automated decision-making (ADM). This statement is required under the Policy for the responsible use of AI in government (the Policy).
How do we define AI & ADM?
In accordance with the Policy, IP Australia applies the definition for AI provided by the Organisation for Economic Co-operation and Development (OECD): An AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment. With reference to the Commonwealth Ombudsman’s Automated Decision-making – Better Practice Guide, IP Australia defines ADM as: Automated Decision-Making refers to the application of automated systems in any part of the decision-making process. Automated decision-making includes using automated systems to: • make the final decision • make interim assessments or decisions leading up to the final decision • recommend a decision to a human decision-maker • guide a human decision-maker through relevant facts, legislation or policy • automate aspects of the fact-finding process which may influence an interim decision or the final decision. Automated systems range from traditional non-technological rules-based systems to specialised technological systems, which use automated tools to predict and deliberate. Why do we use AI & ADM? IP Australia’s purpose is Enabling Australians to benefit from great ideas by providing a world-leading IP system. We are committed to fostering innovation and engaging AI and ADM to improve the efficiency and accessibility of the IP rights system and deliver innovative digital and data-driven customer and staff experiences. How do we use AI & ADM?
IP Australia uses AI and ADM to boost service efficiency by automating business processes, enhancing compliance and fraud detection, and supporting decision-making and analysis of information. Our approach is based on Australia’s AI Ethics Principles and reflects that decisions that are likely to significantly impact our customers, including decisions to grant IP rights, are always made by our skilled people.

2 of 6 Under our IP rights legislation1, the Commissioner of Patents, Registrar of Trade Marks and Designs, and Registrar of Plant Breeder’s Rights may arrange for the use, under their control, of computer programs for any purposes for which they may, or must, under the legislation:
• make a decision;
• exercise any power or comply with any obligation; or
• do anything else related to making a decision, exercising a power, or complying with an obligation. IP Australia currently uses AI in the following domains and usage patterns. More information about AI usage patterns and domains is available on the DTA website: https://www.digital.gov.au/policy/ai/resources/use-classification.

Domains • Service delivery: Enhances the services we provide. AI supports our IP rights examination and administration officers to deliver more efficient services and products, allowing them to focus their expertise on the quality of our products.
• Corporate and enabling: Automates processes, optimises resources and improves efficiency. This includes AI enabled tools that support staff to quickly digest large amounts of information, refine emails and documents, and keep records of meetings.
• Compliance and fraud detection: Identifies patterns or anomalies in data to detect fraudulent activities and support compliance with laws and regulations. • Policy and legal: Analyses policies and legal documents to provide insights that support our expert policy and legal officers. • Scientific: Processes complex datasets, predict outcomes and enhance monitoring functions. Usage patterns • Decision making and administrative action: Supports decision making or the taking of administrative action by guiding, assessing or making a recommendation to a human decision maker.
• Analytics for insights: Identifies, produces or understands insights within structured or unstructured materials via comprehensive data analysis, predictive modelling and/or reporting tools. • Workplace productivity: Automates routine tasks, manages workflows, and facilitates communication. • Image processing: Processes images to identify patterns and objects. Interaction with the public
IP Australia makes specific AI tools available to the public which are designed to assist in accessing information about the IP rights system. This means that the public may directly interact with, or be significantly impacted by, AI or its outputs without human review. This falls under the service delivery domain with usage patterns consistent with analytics for insights, image processing and workplace productivity.

1 Trade Marks Act 1995, Section 222A Computerised decision-making. Designs Act 2003, Section 135A Computerised decision-making. Patents Act 1990, Section 223A Computerised decision-making. Plant Breeder’s Rights Act 1994, Section 76B Computerised decision‑making.

3 of 6 The following are plain English examples of how we use AI to benefit our customers, stakeholders and staff.

Patent examination: Family Member Analyser Domain: Service delivery; Corporate and enabling Usage patterns: Workplace productivity; Decision making and administrative action; Analytics for insights Family members are applications for the same invention lodged in Australia and other countries/jurisdictions. Family Member Analyser is used by IP Australia Patent Examiners to improve the efficiency of the patent examination process. This tool uses a machine learning model to compare published claim sets of ‘family members’ of an Australian Patent Application. The model uses semantic word comparison of independent claims to assess their similarity and then rank the family member search results by claim relevance. This ranking assists patent examiners to quickly identify all available family members and overseas examination reports that are most relevant to the specific case undergoing examination. Family Member Analyser does not make examination decisions, recommendations, or determinations. All examination outcomes remain the responsibility of the Patent Examiner, who exercises independent professional judgement when reviewing and using the information provided by the tool. The ranked results are provided as decision‑support information and may be accepted, supplemented, or disregarded by the examiner as appropriate to the circumstances of the case under Australian law.
Corporate tasks: Microsoft Copilot 365 Domain: Corporate and enabling; Policy and legal Usage patterns: Workplace productivity; Analytics for insights IP Australia staff, across both corporate and examination functions, use Microsoft Copilot 365 to assist with everyday tasks in Microsoft programs like Teams, Outlook, Word, PowerPoint and Excel. Copilot 365 is used by staff in routine tasks such as summarising meetings and chat conversations and assisting in the editing process for documents or emails. The tool is particularly useful to staff in our corporate functions, such as policy and legal, to help get across large or complex document sets. Any outputs generated by Copilot are always subject to the professional expertise and judgement of our staff. TM Checker Domain: Service Delivery
Usage patterns: Analytics for insights; Image Processing TM Checker is a publicly available AI assisted tool provided by IP Australia to help users decide if their trade mark is likely to face problems when it is examined. The tool is designed primarily to support small and medium enterprises by improving access to information about the Australian trade mark system. TM Checker uses AI to check a proposed trade mark based on internally trained data and examinations conducted by IP Australia.
TM Checker provides informational results only and does not make decisions, recommendations, or determinations about registrability. Users should carefully evaluate the accuracy, currency, completeness and relevance of the information obtained from TM Checker and draw their own conclusions for their specific circumstances and purposes. This tool is not a substitute for receiving legal advice in relation to registration for a trade mark. IP Australia recommends that users consult an intellectual property professional directly for specific advice in relation to registering a trade mark.

4 of 6 IP First Response AI Chatbot Domain: Service Delivery
Usage patterns: Analytics for insights The IP First Response AI Chatbot is an experimental tool developed by IP Australia to help individuals and small businesses access plain English information on intellectual property rights and enforcement in Australia. The chatbot provides general guidance and does not offer professional legal advice or recommendations, only options on navigating the IP rights system. The tool can only access and provide users approved content from the publicly viewable webpages on IP First Response website and a curated number of IP Australia corporate webpages – this is its knowledge base. The tool does not access external sources. The chatbot also retains access to the base training data contained within the Claude Large Language Model.
The tool cannot recall conversation history beyond a user’s active session. Any retained data is deidentified before it is stored to help protect user privacy. The deidentified data is used exclusively for monitoring and service improvement. The chatbot has been programmed (via it’s prompts) to restrict discussions to IP rights and IP rights enforcement only, however due to the variable nature of generative AI there may be rare instances where the chatbot could produce outputs outside of the pre-defined context. The deidentified chat data is reviewed to monitor for instances of this context-drift.

Customer Contact Centre’s new systems
Domain: Service delivery Usage patterns: Workplace productivity When a customer contacts IP Australia’s Contact Centre, an AI-powered system called Oration may assist in handling the enquiry. Oration converts speech into text, identifies the reason for the call, and collects information needed to help the customer, such as details about intellectual property (IP) rights. Using pre-defined business rules, the system helps direct the call to the customer service representative best placed to assist. Information from the call is also recorded in Salesforce, our customer relationship management system, to support customer service and record-keeping requirements.
IP Australia also uses AI features within Salesforce, known as Agentforce, to help customer service representatives work more efficiently. These tools can automatically suggest enquiry topics, identify relevant IP right information, summarise customer interactions, and generate draft email or live chat responses. The AI helps reduce administrative tasks and enables staff to focus on providing customer support. AI does not make decisions on behalf of IP Australia or our customers. All AI-generated information and draft responses are reviewed and can be amended by a customer service representative before being saved or sent. AI-generated content is based only on IP Australia-approved knowledge articles, and security controls are in place to protect personal information and ensure customer data remains secure.

5 of 6 ADM use cases At IP Australia, ADM is only used for objective and rules-based decisions and administrative actions, or to support a human decision maker. Decisions that are likely to significantly impact our customers, including decisions to grant IP rights, are always made by our skilled people. The following are plain English examples of how ADM is used to benefit our customers, stakeholders and staff.

Patent administration: Minimum filing requirements
The initial processing for some types of new patent applications has been automated using optical character recognition. The system automatically reviews the application and verifies key elements of the minimum filing requirements, including a language check, applications filed in English, description supplied, applicant/agent contact details. If the automation identifies anything that does not meet minimum requirements, the application is sent to an IP Australia staff member for manual review. If the system identifies the minimum filing requirements are met, the application is sent on for the next step in the process.
Trade Mark administration: Trade marks renewal IP Australia uses basic rules-based automation to manage the renewal process for registered trade marks. When a trade mark is coming up for renewal, which happens every 10 years, the system automatically notifies the trade mark owner. The trade mark owner can pay the renewal fee any time in the 12 months before it’s due, or up to 6 months after the due date. If the fee is paid, the system automatically renews the trade mark. If the fee isn’t paid within those 6 months after the deadline, the system will automatically mark the trade mark as ceased. Automated reminders and notifications are sent throughout the process, making it easier for trade mark owners to keep their registration active without missing any deadlines, and generating significant efficiencies for IP Australia in trade mark administration.
How do we govern AI & ADM? IP Australia has AI and ADM governance policies. These policies set the guardrails to ensure appropriate governance of AI and ADM technologies at IP Australia. While distinct, the two polices share many features and are governed in similar ways. We review the policies at least annually to ensure they continue to be appropriate for IP Australia’s context and are aligned to
whole-of-government requirements.
IP Australia’s Accountable Official for AI and Chief AI Officer is the Deputy Director General. The Accountable Official is accountable for governance of AI and implementation of the Policy within IP Australia.
As per the IP rights legislation, the Commissioner/Registrar is responsible for all relevant applications of ADM.
To ensure appropriate monitoring and governance of AI, IP Australia has implemented the Australian Government Artificial intelligence impact assessment tool, and applies additional risk-scaled governance mechanisms. This includes regular review of deployed AI systems.
Our approach to governance empowers our people to engage in low-risk AI use cases and ensures that higher-risk applications are subject to rigorous controls and conditions to protect the public, our customers and our staff.

6 of 6 How do we comply with the Policy? IP Australia is compliant with all mandatory requirements of the Policy, except for those mandatory requirements with a future implementation date articulated in the Policy, which we are working towards applying.
When will this transparency statement be updated? This statement will be updated at least once a year, or sooner if we make any significant change to our approach to AI, or if any new factor materially impacts the existing statement’s accuracy. Who can I talk to? For any enquires regarding IP Australia’s use of AI, please contact
MDB-AI-ADM-Governance@IPAustralia.gov.au.

Version History
Version number
Date published Key changes
1 27 February 2025 First version
2 27 February 2026 Annual update
3 30 July 2026 Interim Update: Introduction of the IP First Response AI Chatbot and Customer Contact Centre’s new systems