QUESTION Could the Minister please advise on the following: Is the ATO using any artificial intelligence [AI] system/s to assess data, submissions or returns from the public and organisations, and if so: please detail all systems in use including the architecture/technique and functional role of the AI systems in use, for example large language models and/or generative AI; and where is the data/information used by the AI systems stored, is that information held onshore or in servers overseas. Can the ATO advise what human oversight these systems have; and of these people, w hat are their qualifications and experience. How has the ATO informed Australian citizens that their data, information and submissions are being assessed by AI tools. How often does the ATO run quality assurance checks on the systems. Has the ATO coordinated with other departments on the deployment of AI; and if so, which ones. What checks and balances are in place to ensure that the AI systems are operating fairly and effectively. ANSWER Senator the Hon Katy Gallagher: The Treasurer - The Hon Jim Chalmers MP has provided the following answer to the Honourable Senator's question as follows: Yes Machine Learning, Natural Language Processing and Large Language Models. The ATO uses AI to assist staff to quickly process large volumes of complex data to identify insights and risks. The AI used helps ATO staff make more informed decisions. It frees them to focus more on aspects of work that require human judgment and empathy. The ATO also uses AI to ensure we can continue to protect the tax and superannuation systems from external fraud and cyber attacks. Data is stored within the ATO's secure IT environments, including on premise data centre tenancies, and virtual private cloud tenancies with AWS and Azure, all of which apply appropriate enterprise data protection. They are all on shore and controlled through vendor contractual obligations and verified through IT security assessments. All ATO staff are required to follow the ATO's AI Senior Executive Instruction (SEI) which sets out their responsibilities when interacting with artificial intelligence (AI), using AI outputs and developing AI systems. Under the SEI, staff are required to verify all AI outputs for accuracy before they are used to make a decision, deliver services to clients, or as final products. AI use is also limited to approved tools for approved use cases. Specialised roles outlined in the SEI include: AI Accountable Official - Chief Data Officer, Deputy Commissioner, Smarter Data AI Lifecycle Subject Matter Expert - Assistant Commissioner, Data Management, Smarter Data AI Risk Owner - Assistant Commissioner, Data Science, Smarter Data AI Use Case Owner - Deputy Commissioner or Assistant Commissioner Data and Analytics Assurance Committee - chaired by the Chief Data Officer, Deputy Commissioner, Smarter Data The ATO has highly trained and skilled data scientists and data engineers who develop and deploy our in-house enterprise AI tools and models under strictly controlled IT deployment processes. These data scientists and data engineers sit in our Smarter Data business line. Taxpayers have a right of review of decisions that impact them. Their rights apply regardless of the technology we use to help us make those decisions. The ATO publishes an AI Transparency Statement on their website which outlines to the community how they approach AI in the ATO. This is available at https://www.ato.gov.au/about-ato/commitments-and-reporting/information-and-privacy/ato-aitransparency-statement . ATO staff are required to follow quality assurance processes to verify accuracy of AI outputs before they are used to make a decision or deliver a service to clients. The AI SEI requires staff to review the risk of their AI systems annually, or whenever their data inputs, outputs, or use case materially changes. ATO's policies and guidelines, including the AI Senior Executive Instruction and deployment of AI aligns with APS guidance provided by the Digital Transformation Office, as part of the APS AI Plan. The ATO has signed up to the GovAI pilot that is led by the Department of Finance and is exploring potential AI models using synthetic data in a sandbox environment. The ATO takes a measured and cautious approach when exploring and deploying AI tools and techniques The ATO closely monitor, and take appropriate action, if intended benefits are not being realised or unintended consequences arise, and actively identify and manage risks associated with all applications of AI. The ATO only uses AI in ways that maintain the security, privacy, transparency and ethical use of data as well as always maintain appropriate human oversight. The AI SEI requires ATO staff to assess and manage impact and fairness when developing AI solutions.