Published on the 05/02/2025 | Written by Heather Wright

Private sector to follow…
The New Zealand government has released guidelines for the responsible adoption of AI in the public sector, with work on similar guidance for the private sector currently underway.
An initial ‘plan on a page’ framework, lacking any real detail but big on responsible AI buzzwords, was released last week by the Department of Internal Affairs and followed by more detailed guidance this week setting expectations for how agencies should adopt AI.
“AI systems are evolving rapidly, and government policies, guidance and use cases will continue to adapt.”
Last year the government announced it would take a ‘light touch, proportionate and risk-based’ approach to AI regulation.
The Public Service Artificial Intelligence Framework was led by the Government Chief Digital Officer (GCDO) and is based on the OECD’s AI principles, which the government committed to adopting in June 2024 to bring New Zealand in line with other OECD member states.
It sets out a structured approach to safely deploy all forms of AI in the public sector. Among the principles is that agencies must publicly disclose how the AI used affects any outcomes.
The framework includes six pillars:
- Governance – ensuring human accountability for inclusive implementation of data and AI use
- Guardrails – supporting safe and trustworthy use of AI and its underpinning data
- Capability – building internal and external AI capability and safety by design
- Innovation – developing pathways for safe AI innovation at government agencies
- Social license – understanding how to build public trust and worker engagement
- Global voice – building New Zealand’s global reputation as a trusted partner and public service AI enabler.
Paul James, Department of Internal Affairs chief executive and Government Chief Digital Officer, says the framework applies to all forms of AI and sets out a structured approach to ensure safety and privacy remain central to AI use, while helping agencies boost efficiency and deliver better outcomes and building AI capability across the public service.
The framework was quickly followed by updated guidance on the use of generative AI with the Responsible AI Guidance for the Public Sector: GenAI guidelines, which sit below the framework. Both sit within the wider National AI Strategy, which is still in development.
In releasing the guidance, Judith Collins, Minister for Digitising Government, emphasised the ‘major opportunity’ presented by AI to lift productivity and improve public service delivery.
“Use of AI technologies to improve public services is a priority to me,” Collins says.
“Harnessing AI effectively can significantly improve customer experience and boost efficiency.
“It can help reduce wait times, triage issues faster and allow public servants to focus on frontline services – and that means delivering better outcomes for New Zealanders while reducing costs to government.”
The Responsible AI Guidance says it is designed to support the public service to explore generative AI systems in ways that are ‘safe, transparent and responsible’.
A 2024 cross-agency AI survey by the GCDO found 37 out of the 50 agencies surveyed had at least one AI use cases, with 108 total use cases from the 37 agencies.
The majority of the use cases identified were focused on internal agency use for improving productivity and efficiency, including chatbots to help people find more appropriate guidance and resources, document drafting and meeting summaries; automated data summary and analysis; and automated transcription.
The survey also highlighted the use of digital detection and analysis AI tools, providing analysis and detection for large image datasets as well as retinal scanning.
Customer-focused use cases included self-service with automated and predictive forms, and assisted web search using chatbots to help customers find information more quickly and easily.
But the survey also noted risks and barriers to AI adoption for the public sector, with skills and capabilities issues leading the way, followed by policies, guidance and privacy and security concerns.
The newly released guidance is part of a suite of tools for agencies to adopt AI in ways that are safe, transparent and deliver ‘real value’ for New Zealanders, while upholding the highest standards of trust, Collins says.
The GCDO is working with the Ministry of Business Innovation and Employment to develop similar guidance for the business community, she says.
The New Zealand Privacy Commissioner released guidance of AI in relation to personal information and privacy aspects, back in 2023.
Agencies have joined up to support responsible AI adoption across both government and industry.
“AI systems are evolving rapidly, and government policies, guidance and use cases will continue to adapt alongside these advancements and public expectations,” Collins says.
In Australia, the Digital Transformation Agency introduced a policy for the responsible use of AI in government, which came into effect in September.
It is driven by an ‘enable, engage and evolve’ framework to introduce principles, mandatory requirements and recommended actions, and included the requirement that every agency identified accountable officials and ‘provided them to the DTA’ within 90 days of the policy taking effect.