Skip to content

We’ve seen a lot of interest in Artificial Intelligence (AI) in recent years, especially generative AI. Here are some good pieces of AI guidance you may like to consider for your project.

  • Responsible AI Guidance for the Public Service: GenAI supports the New Zealand Public Service to explore generative artificial intelligence (GenAI) systems in ways that are safe, transparent and responsible. It updates and expands on the ‘Interim Generative AI guidance for the public service’ published in July 2023 in response to feedback from agencies.
  • The Office of the Privacy Commission released guidance on Artificial Intelligence and the information privacy principles.
  • The AI Forum has developed a website of AI Governance which offers resources for inclusive and responsible AI adoption, including working with Māori data along with practical tool kits to support you in your AI journey.
  • The OCED Principles for trustworthy AI offer value-based principles, with examples of the principles in practice.
  • The International Science Council, led by Sir Prof Peter Gluckman, released A guide for policy-makers: Evaluating rapidly developing technologies including AI, large language models and beyond in April 2024. The guide comes with a useful framework you can fill out with questions to help you evaluate a large language model.
  • If you are looking at procuring AI, the Algorithm Impact Assessment Questionnaire offers guidance about procuring algorithms. It is a good place to start thinking about the issues you’ll need to consider. Remember, there are a number of steps in procuring AI where you’ll need to think about ethics. This includes:
    • Pre-procurement - is AI really the solution? Would another tested technology be just as good for your problem?
    • During the procurement process – do you understand the training data? Who owns the data the system has been trained on? Does the data fit our New Zealand context?
    • Post-procurement – even when the AI is embedded, you will need to continue to monitor and test the system to make sure it has not drifted, and that bias is not present.
Top