QUESTION Could the Minister please advise on the following: What artificial intelligence [AI] system/s are being used to assess submissions from the public and organisations, and in relation to these: please detail all systems in use including the architecture/technique and functional role of the AI systems in use, for example large language models or generative AI; and where is this data/information stored; is it retained onshore or held in servers overseas. What has the Department done to inform Australian citizens that their submissions are being assessed by AI. Are AI models being used to assess applications for Department grants, for example Australian Industry Participation, CAIP and the Industry Growth Program. What checks are in place to ensure grants are being assessed appropriately. Can the Department define what type of AI systems are in use (for example algorithmic systems or advanced automation). Do these systems have human oversight, and if so: how many people; and please detail their qualifications and experience. Does the Department coordinate with other relevant departments on the deployment of AI. How often does the Department run quality assurance checks on these systems. What checks and balances are in place to ensure that the AI assessment model employed by the Department is operating fairly and effectively. Has the Department investigated the developments in, or viability of, smaller, targeted AI language models; and if not, why not. ANSWER Senator Ayres: Please see my answer below to the Senator's question: Use of artificial intelligence in the department The Department of Industry Science and Resources' (the department) Artificial Intelligence Transparency Statement ( https://www.industry.gov.au/publications/artificial-intelligence-transparency-statement ) is a useful guide for general details on how the department uses artificial intelligence (AI). The department has used AI-powered public consultation summary analysis to process large-scale public consultation submissions. This tool helps staff summarise submissions, identify recurring themes, and group responses for further human review. All outputs are treated as analytical support and reviewed by people. The department informs submitters about the use of AI at the point of consultation through privacy collection statements, the department's privacy policy, and its AI Transparency Statement. Where AI-enabled tools are used to support analysis, this is clearly disclosed and only undertaken where informed consent is provided in accordance with privacy requirements. The department does not train AI models on public submissions or other departmental data. AI tools are used to generate draft outputs that support staff analysis, not to automate assessments or decisions. All data processed through these systems is retained within the department's Microsoft Azure tenancy and stored in Australian based data centres, consistent with Commonwealth data residency, security and privacy requirements. Human oversight is applied throughout all stages of AI use to ensure outputs are appropriate, accurate and consistent with policy, legal and ethical obligations. The department monitors developments in AI technologies, including smaller or more targeted language models, as part of its broader assessment of emerging capabilities. Technology choices are guided by fitness for purpose, security, privacy, assurance requirements and alignment with whole-of-government governance frameworks. The department will continue to consider alternative models where they are appropriate, proportionate and demonstrably safe. The department does not use AI models to assess applications for any department grants, including Australian Industry Participation, CAIP and the Industry Growth Program.