Tuanz urges trust rules for New Zealand generative AI

TUANZ has released a policy paper on responsible generative AI in New Zealand, arguing that trust will shape how widely the technology is adopted.

The paper is part of TUANZ's policy programme and identifies four priorities as businesses and public agencies expand their use of generative AI: investment in domestic AI work, explainability rules for high-stakes decisions, a national standard for human oversight, and practical support for small and medium-sized enterprises.

TUANZ says AI could contribute as much as $76 billion to New Zealand's economy by 2038, but argues that public confidence will determine whether that potential is realised. Concerns about fairness, transparency and data use are still affecting willingness to adopt AI tools and services.

It also points to a gap between early experimentation and broader practical use, especially among smaller firms. Many SMEs remain at the stage of awareness rather than implementation, raising the risk that larger organisations move ahead while smaller businesses fall behind.

Chair Paul Littlefair linked the group's recommendations to productivity and economic performance.

"AI presents a significant opportunity for New Zealand to lift productivity and strengthen our global competitiveness," Littlefair said.

"Adoption is already underway, but trust will determine how far and how fast we can go."

Policy focus

One of TUANZ's main proposals is stronger public backing for AI work based in New Zealand. It says government should continue to support international technology investment while also funding local AI services and research so the domestic sector can remain internationally competitive.

Another recommendation focuses on what it calls algorithmic explainability in high-stakes settings. Organisations using AI to make life-affecting decisions in areas such as health or banking should be required to provide clear explanations of how those decisions are reached, rather than leaving individuals to deal with opaque systems.

TUANZ also wants a national human-in-the-loop standard. It argues that people are more comfortable using AI when they know a person can intervene, and that industry standards should require escalation routes to human support when systems produce biased or inaccurate outputs.

A fourth proposal is the creation of an AI readiness resource hub for SMEs. TUANZ says it would work with government and other groups on tools and webinars designed to help smaller businesses move from uncertainty to practical use of the technology.

Trust question

TUANZ argues that New Zealand should seek an advantage through the way it governs and applies AI rather than through sheer size. The position reflects a broader debate across many markets over whether public trust, accountability and visibility into automated systems will become as important as technical adoption rates.

Littlefair framed that as a matter of national positioning.

"We have an opportunity to lead not through scale, but through trust," he said.

"That means ensuring AI systems are transparent, accountable and designed around people."

TUANZ says confidence in AI affects not only consumer acceptance but also whether organisations choose to spend on deployment and integration. In its view, responsible AI use is tied to broader economic participation and long-term growth, rather than being only a question of risk management.

The proposals come as organisations in New Zealand and elsewhere weigh how to use generative AI tools without losing oversight of decisions that affect customers, patients and workers. Questions of explainability, accountability and human review have become central in sectors where automated recommendations may influence lending, healthcare and other important outcomes.

For smaller businesses, the challenge is often less about principle than access to practical guidance, internal expertise and clear examples of compliant use. TUANZ's focus on SME support suggests it sees the next stage of adoption as depending on whether smaller firms can move beyond experimentation and understand where human review and disclosure are needed.

Littlefair said trust will determine whether people and companies commit more fully to the technology.

"Trust is fundamental to the success of AI," he said.

"When people and businesses feel confident in how these systems operate, they are more likely to adopt them, invest in them and realise their full value.

"Getting this right positions New Zealand to capture the benefits of AI in a way that is inclusive, credible and aligned with our values."