Glossary

Key terms used throughout this document:

AGI (Artificial General Intelligence) — A hypothetical AI system with human-level cognitive abilities across all domains. Distinct from current "narrow" AI systems that excel at specific tasks.

AI (Artificial Intelligence) — Computer systems designed to perform tasks that typically require human intelligence, including learning, reasoning, and problem-solving.

Algorithmic accountability — The principle that organisations using AI to make decisions should be transparent about how those decisions are made and responsible for their outcomes.

Automation — The use of technology to perform tasks without human intervention. In the AI context, this increasingly includes cognitive tasks that were previously considered resistant to automation.

Data sovereignty — The principle that data is subject to the laws and governance of the nation where it is collected.

Generative AI — AI systems that can create new content (text, images, code, music) based on patterns learned from training data.

Hauora — Maori concept of health and wellbeing, encompassing physical, mental, spiritual, and family/community dimensions.

Kaitiakitanga — Guardianship and stewardship — a Maori concept relevant to how we manage AI's impact on people and environment.

LLM (Large Language Model) — AI systems trained on vast amounts of text data, capable of generating and understanding natural language. Examples include GPT, Claude, and Gemini.

Te Mana Raraunga — Maori data sovereignty. The principle that Maori data should be subject to Maori governance.

Te Tiriti o Waitangi — The Treaty of Waitangi. New Zealand's founding document, establishing the relationship between the Crown and Maori, with ongoing implications for AI governance and data rights.

UBI (Universal Basic Income) — A regular cash payment to all citizens or residents, without means testing or work requirements.

Algorithmic management — Where software systems direct, monitor, and evaluate workers in real time — setting prices, assigning work, and making employment decisions with minimal human involvement. Already present in NZ logistics, retail, and gig economy platforms.

CLOUD Act — US legislation (2018) allowing US law enforcement to compel disclosure of data stored anywhere in the world by US-incorporated providers, regardless of where the data physically sits. Relevant because most NZ government cloud workloads run on AWS Sydney or Azure Australia.

Flexicurity — A labour market model combining easy hiring and firing with generous unemployment benefits and intensive retraining. Denmark's version provides up to 90% of previous income for two years. Referenced as a model for NZ transition policy.

Micro-credentials — Short, assessed, NZQA-recognised qualifications in specific competencies. A potentially powerful instrument for rapidly upskilling workers in AI-adjacent skills. The framework exists in NZ; the question is whether employers and immigration systems recognise them equivalently.

NCEA (National Certificate of Educational Achievement) — New Zealand's main secondary school qualification system. Has more flexibility for oral and practical assessment than many international equivalents — relevant as AI makes traditional written exams harder to administer meaningfully.

Negative income tax — A system where the government sets an income threshold below which it makes tax credit payments that shrink as income rises. Proposed by Milton Friedman in 1962. NZ's Working for Families is a partial version.

Open-weight models — AI models (such as Meta's Llama, Mistral, DeepSeek) whose parameters are publicly released, allowing anyone to run, fine-tune, and deploy them. Distinct from proprietary API-only models. The basis for NZ sovereign inference proposals.

Sinking lid — The dominant mechanism of AI-driven job displacement in NZ: when people leave, they are not replaced. 40% of NZ companies report reduced need for new hires, but only 7% report directly replacing workers. The effect is the same — fewer jobs — but it happens quietly, one unfilled vacancy at a time.

Sovereign inference — Government-owned AI compute infrastructure running open-weight models on NZ soil, ensuring sensitive workloads are processed on systems NZ controls. Estimated capital cost of NZD $3–5 million for a government-scale cluster.

Te Pūkenga — New Zealand's national vocational education network, formed by the 2020 merger of 16 polytechnics and 9 industry training organisations. Now being restructured. The system's state of active redesign creates both opportunity and risk for embedding AI literacy.

Whanaungatanga — The Maori concept of kinship, connection, and belonging — relevant to community-based approaches to transition support.