38367 (2024). Francisco Hernandez to the Minister for Biosecurity

38367 (2024). Francisco Hernandez to the Minister for Biosecurity: How, if at all, are the Minister's departments, if any, currently using AI, and what plans, if any, have been developed to manage the risks associated with AI?

Hon Andrew Hoggard: I am advised that the Ministry for Primary Industries (MPI) has very limited use of Artificial Intelligence (AI) primarily using Machine Learning. These are listed below: • helping classify images as part of our standard forestry geospatial technology; • helping to identify video footage of interest for human analysis in Fisheries; • to automatically blur identifying human features from camera footage in Fisheries; • helps improve readability and style checking of content for the MPI website; • assists us to analyse and produce reports on the way the MPI website is being used; and • for diagnostics optimisation in Biosecurity assay design. These have been through MPI’s standard IT architecture and information security governance and assurance processes. Recognising the significant opportunities and the need to balance these against the potential risks of AI, MPI established the Artificial Intelligence Technical Advisory Group (AITAG) in late 2023. The group’s purpose is to manage risks and promote safe adoption of AI through draft policy, guidance, and processes that will inform MPI’s on-going safe use of artificial intelligence. The AITAG is in the process of developing a policy for the use of artificial intelligence in the workplace. The policy will reference principles included in the current draft Artificial Intelligence Framework for the New Zealand Public Service produced by the Government Chief Digital Officer. The AITAG is also compiling a register of current and proposed AI use across the organisation as part of its governance process. MPI’s approach to AI adoption will be measured, focusing initially on a very limited number of AI use case opportunities where it could deliver efficiencies, cost savings, or improve the productivity of MPI operations.