Article: Polyanna policy – Is NZ’s framework for AI use in government overly optimistic?

11/05/2026

‘Polyanna policy’: Is the framework for AI use in government overly optimistic? (2026)

In this 2026 piece for The Conversation, Barbara Allen and I ask whether the governance frameworks Aotearoa is building to manage artificial intelligence can be trusted to work, at a time when public service agencies are being encouraged to adopt it. We suggest that the Public Service AI Framework names the right principles, transparency, fairness and human oversight, but that because it is explicitly non-binding it leans towards optimism rather than accountability, which is why we call it a Pollyanna policy. We argue that the framework tends to underestimate the layered and often conflicting obligations inside public agencies, and that leaving compliance to agencies with very different levels of capability shifts accountability away from the centre. The 2025 Public Service Census, which found that a third of public servants had used AI for work but only 14 percent used it regularly, suggests that organisational readiness cannot be assumed. For Māori data sovereignty, an obligation under Te Tiriti o Waitangi, we see a particular risk that fabricated or sycophantic outputs will overwrite Indigenous knowledge. We conclude that public institutions need formal, protected roles in which officials test AI output for bias and fabrication, rather than treating speed as a measure of quality.

Read the article: https://theconversation.com/polyanna-policy-is-nzs-framework-for-ai-use-in-government-overly-optimistic-281425

Reference
Te Kawa, D., & Allen, B. (2026, May 11). ‘Polyanna policy’: Is NZ’s framework for AI use in government overly optimistic? The Conversation. https://theconversation.com/polyanna-policy-is-nzs-framework-for-ai-use-in-government-overly-optimistic-281425

ORCID: https://orcid.org/0009-0006-5129-5218