---
title: "How AI informs New Zealand voters"
publisher: "Onsomble Limited"
publication: "NZBrands.ai"
edition: 3
collected_on: "2026-10-16"
html_url: "https://nzbrands.ai/research/nz-election-2026/editions/3/"
---

# How AI informs New Zealand voters

Edition 3 · answers collected 16 October 2026.

[Read on NZBrands.ai](https://nzbrands.ai/research/nz-election-2026/editions/3/)

A study of what AI assistants tell New Zealanders about the 2026 general election. We put 174 questions, written as 32 New Zealand voters, to ChatGPT, Gemini, Google AI Mode and Google AI Overviews, and analysed the responses: which parties each one named, what it said about them, and which sources it relied on.

## Study figures

| Measure | Value | Meaning |
| --- | --- | --- |
| AI assistants | 4 | The four AI tools we asked: ChatGPT, Gemini, Google AI Mode, and the AI Overviews that appear at the top of Google search results. |
| questions | 174 | The questions we put to every assistant, each written the way a real voter would ask it. |
| voter profiles | 32 | Every question was asked as one of 32 New Zealanders, such as a first-time voter, a sole parent or a farming household, so the answers reflect different people’s lives. |
| topics | 20 | The subjects the questions cover, such as housing, healthcare, retirement and infrastructure. |
| locations | 9 | We also asked from nine locations around the country, from Northland to Queenstown Lakes, to see whether the answers change depending on where you are. |
| answers analysed | 793 | The 793 answers produced on 16 October. Google did not produce an AI Overview for 27 of the planned question-and-location combinations. |
| last updated | 20 Oct | Party-identity coding corrected on 20 October. This edition contains the answers collected on 16 October. |

## Key results at a glance

### Ask the same question again and you will probably hear about different parties.

531 of 792 — matched answers changed which of the 17 tracked parties they named

Across all 17 tracked parties, 531 of 792 matched answers changed their party line-up. The seven-party charts below show a narrower set. Coding for both editions was corrected on 20 October; the original answers are unchanged.

[See which parties were named](https://nzbrands.ai/research/nz-election-2026/editions/3/#data)

### TOP is turning up far more often.

246 of 792 — answers that mentioned The Opportunity Party

On 12 October, one answer in five mentioned The Opportunity Party. Four days later it was nearly one in three. The six parties in Parliament moved much less.

[Compare the parties](https://nzbrands.ai/research/nz-election-2026/editions/3/#data)

### National and location answers often named different parties.

61 of 123 — answers that changed when asked from a location

We compared the same questions nationally and from nine locations. Different parties were named in 61 of 123 complete pairs. These captures show variation; they do not establish that location caused the difference.

[Explore the nine locations](https://nzbrands.ai/research/nz-election-2026/editions/3/#situations)

### AI will not tell you who to vote for, but it will point you in a direction.

22 of 31 — answers that pointed the voter toward particular parties

When voters asked who to vote for, no answer named a single party. But 22 of the 31 answers said which parties fit the voter’s priorities, up from 12 in the last edition.

[See what the answers did](https://nzbrands.ai/research/nz-election-2026/editions/3/#guidance)

### More Google AI Overviews now show where their information came from.

62 of 177 — Google AI Overviews that included a link

On the same 177 questions, the number of AI Overviews that gave a link you could click rose from 51 to 62.

[See links by assistant](https://nzbrands.ai/research/nz-election-2026/editions/3/#sources)

### Two of the twelve facts we checked were wrong.

2 of 12 — checked statements that were wrong

Gemini said party lists are secret, when the Electoral Commission publishes them, and gave overseas voters the wrong deadline for returning a paper vote. Four more statements were unclear or left out an important condition.

[See every check](https://nzbrands.ai/research/nz-election-2026/editions/3/#checks)

## Most answers leave out at least one party in Parliament.

The charts show seven parties: the six parties in Parliament and Opportunity. Of the 723 matched answers that mentioned a parliamentary party, 449 left at least one out. The headline change count above covers all 17 tracked parties.

### Seven displayed parties: how often each was named

| Party | Answers naming party (of 792) | 5 October: answers naming party | Change in answers |
| --- | --- | --- | --- |
| National | 720 | 709 | 11 |
| Labour | 705 | 699 | 6 |
| Greens | 687 | 677 | 10 |
| ACT | 617 | 605 | 12 |
| NZ First | 537 | 511 | 26 |
| Te Pāti Māori | 342 | 302 | 40 |
| Opportunity | 246 | 166 | 80 |

These seven parties are the six parliamentary parties plus Opportunity. Each party counts once per answer. The upper bar is 16 October; the lower bar is 12 October, restated under the same identity corrections. Both use 792 matched answers. The headline covers all 17 tracked parties.

### How often each AI assistant named every party in Parliament

| Assistant | Answers naming all six parliamentary parties | Answers collected |
| --- | --- | --- |
| ChatGPT | 95 | 177 |
| Gemini | 68 | 177 |
| Google AI Mode | 37 | 177 |
| Google AI Overviews | 28 | 177 |

Each square is one answer. Raised squares named all six parties in Parliament. Changes are percentage points on the same 177 cells per assistant.

## Seven displayed parties, issue by issue

The tables show all four assistants together. Answer counts and claim counts are distinct measures; each claim counts only for the party it is about.

## Housing

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 96 | 98 | 98.0% | 139 | 309 | 59 | 29 |
| Labour | 95 | 98 | 96.9% | 160 | 283 | 26 | 32 |
| Greens | 93 | 98 | 94.9% | 205 | 304 | 11 | 13 |
| ACT | 86 | 98 | 87.8% | 98 | 231 | 35 | 15 |
| NZ First | 69 | 98 | 70.4% | 40 | 145 | 23 | 12 |
| Te Pāti Māori | 39 | 98 | 39.8% | 37 | 84 | 7 | 5 |
| Opportunity | 42 | 98 | 42.9% | 35 | 84 | 6 | 2 |

## Infrastructure and local services

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 75 | 95 | 78.9% | 102 | 330 | 50 | 26 |
| Labour | 68 | 95 | 71.6% | 109 | 199 | 21 | 19 |
| Greens | 68 | 95 | 71.6% | 116 | 239 | 9 | 13 |
| ACT | 52 | 95 | 54.7% | 38 | 185 | 25 | 13 |
| NZ First | 52 | 95 | 54.7% | 25 | 163 | 23 | 8 |
| Te Pāti Māori | 18 | 95 | 18.9% | 8 | 48 | 3 | 1 |
| Opportunity | 18 | 95 | 18.9% | 15 | 38 | 3 | 4 |

## Healthcare

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 69 | 77 | 89.6% | 97 | 266 | 18 | 16 |
| Labour | 69 | 77 | 89.6% | 202 | 212 | 11 | 13 |
| Greens | 66 | 77 | 85.7% | 148 | 191 | 0 | 7 |
| ACT | 58 | 77 | 75.3% | 45 | 201 | 4 | 13 |
| NZ First | 56 | 77 | 72.7% | 49 | 159 | 8 | 9 |
| Te Pāti Māori | 42 | 77 | 54.5% | 50 | 95 | 3 | 5 |
| Opportunity | 29 | 77 | 37.7% | 19 | 60 | 1 | 3 |

## Cost of living

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 68 | 70 | 97.1% | 105 | 232 | 23 | 19 |
| Labour | 69 | 70 | 98.6% | 174 | 234 | 9 | 16 |
| Greens | 67 | 70 | 95.7% | 130 | 206 | 6 | 16 |
| ACT | 59 | 70 | 84.3% | 52 | 166 | 20 | 8 |
| NZ First | 49 | 70 | 70.0% | 54 | 135 | 11 | 8 |
| Te Pāti Māori | 42 | 70 | 60.0% | 42 | 99 | 3 | 5 |
| Opportunity | 18 | 70 | 25.7% | 7 | 23 | 2 | 2 |

## Jobs and unemployment

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 58 | 58 | 100.0% | 110 | 249 | 35 | 19 |
| Labour | 58 | 58 | 100.0% | 164 | 233 | 18 | 23 |
| Greens | 53 | 58 | 91.4% | 100 | 200 | 16 | 19 |
| ACT | 49 | 58 | 84.5% | 67 | 190 | 26 | 13 |
| NZ First | 45 | 58 | 77.6% | 48 | 146 | 20 | 9 |
| Te Pāti Māori | 25 | 58 | 43.1% | 24 | 76 | 6 | 5 |
| Opportunity | 13 | 58 | 22.4% | 14 | 40 | 4 | 0 |

## Education and childcare

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 54 | 55 | 98.2% | 94 | 228 | 25 | 12 |
| Labour | 53 | 55 | 96.4% | 128 | 187 | 5 | 15 |
| Greens | 54 | 55 | 98.2% | 137 | 139 | 5 | 6 |
| ACT | 46 | 55 | 83.6% | 33 | 161 | 10 | 7 |
| NZ First | 35 | 55 | 63.6% | 24 | 105 | 11 | 6 |
| Te Pāti Māori | 39 | 55 | 70.9% | 58 | 97 | 5 | 1 |
| Opportunity | 14 | 55 | 25.5% | 17 | 24 | 3 | 2 |

## Tax and public spending

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 50 | 55 | 90.9% | 69 | 169 | 28 | 22 |
| Labour | 47 | 55 | 85.5% | 103 | 133 | 17 | 25 |
| Greens | 46 | 55 | 83.6% | 82 | 131 | 10 | 24 |
| ACT | 37 | 55 | 67.3% | 35 | 114 | 16 | 10 |
| NZ First | 25 | 55 | 45.5% | 28 | 56 | 8 | 7 |
| Te Pāti Māori | 19 | 55 | 34.5% | 21 | 51 | 2 | 6 |
| Opportunity | 19 | 55 | 34.5% | 18 | 54 | 1 | 5 |

## Democratic institutions and rights

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 20 | 51 | 39.2% | 8 | 41 | 6 | 1 |
| Labour | 16 | 51 | 31.4% | 10 | 42 | 0 | 0 |
| Greens | 14 | 51 | 27.5% | 16 | 42 | 0 | 0 |
| ACT | 14 | 51 | 27.5% | 0 | 43 | 3 | 3 |
| NZ First | 9 | 51 | 17.6% | 0 | 26 | 2 | 0 |
| Te Pāti Māori | 9 | 51 | 17.6% | 6 | 23 | 0 | 1 |
| Opportunity | 8 | 51 | 15.7% | 5 | 19 | 0 | 0 |

## Fuel and energy

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 35 | 35 | 100.0% | 54 | 157 | 8 | 8 |
| Labour | 35 | 35 | 100.0% | 60 | 143 | 13 | 8 |
| Greens | 35 | 35 | 100.0% | 49 | 150 | 11 | 14 |
| ACT | 34 | 35 | 97.1% | 44 | 132 | 9 | 6 |
| NZ First | 34 | 35 | 97.1% | 39 | 129 | 4 | 5 |
| Te Pāti Māori | 12 | 35 | 34.3% | 3 | 22 | 3 | 2 |
| Opportunity | 9 | 35 | 25.7% | 4 | 16 | 1 | 1 |

## Economy

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 31 | 31 | 100.0% | 70 | 111 | 13 | 6 |
| Labour | 31 | 31 | 100.0% | 63 | 119 | 15 | 7 |
| Greens | 29 | 31 | 93.5% | 37 | 86 | 13 | 5 |
| ACT | 29 | 31 | 93.5% | 49 | 91 | 8 | 5 |
| NZ First | 23 | 31 | 74.2% | 21 | 75 | 8 | 2 |
| Te Pāti Māori | 10 | 31 | 32.3% | 3 | 29 | 5 | 0 |
| Opportunity | 5 | 31 | 16.1% | 2 | 17 | 2 | 0 |

## Poverty and welfare

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 30 | 31 | 96.8% | 27 | 117 | 16 | 7 |
| Labour | 30 | 31 | 96.8% | 71 | 78 | 5 | 6 |
| Greens | 30 | 31 | 96.8% | 89 | 72 | 1 | 1 |
| ACT | 29 | 31 | 93.5% | 18 | 99 | 15 | 5 |
| NZ First | 20 | 31 | 64.5% | 16 | 37 | 9 | 4 |
| Te Pāti Māori | 19 | 31 | 61.3% | 24 | 44 | 3 | 1 |
| Opportunity | 18 | 31 | 58.1% | 27 | 21 | 3 | 0 |

## Climate and disaster resilience

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 20 | 20 | 100.0% | 23 | 84 | 14 | 7 |
| Labour | 20 | 20 | 100.0% | 31 | 91 | 3 | 3 |
| Greens | 20 | 20 | 100.0% | 23 | 106 | 1 | 6 |
| ACT | 20 | 20 | 100.0% | 12 | 88 | 8 | 2 |
| NZ First | 18 | 20 | 90.0% | 6 | 76 | 5 | 2 |
| Te Pāti Māori | 9 | 20 | 45.0% | 6 | 36 | 1 | 1 |
| Opportunity | 10 | 20 | 50.0% | 8 | 37 | 3 | 3 |

## Retirement and aged care

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 18 | 19 | 94.7% | 16 | 84 | 5 | 1 |
| Labour | 18 | 19 | 94.7% | 31 | 74 | 1 | 1 |
| Greens | 18 | 19 | 94.7% | 22 | 66 | 0 | 2 |
| ACT | 14 | 19 | 73.7% | 4 | 56 | 1 | 2 |
| NZ First | 18 | 19 | 94.7% | 8 | 75 | 3 | 1 |
| Te Pāti Māori | 10 | 19 | 52.6% | 5 | 30 | 1 | 0 |
| Opportunity | 11 | 19 | 57.9% | 6 | 51 | 1 | 1 |

## Rural and agricultural policy

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 19 | 19 | 100.0% | 49 | 119 | 11 | 4 |
| Labour | 19 | 19 | 100.0% | 20 | 110 | 13 | 6 |
| Greens | 19 | 19 | 100.0% | 17 | 112 | 15 | 5 |
| ACT | 19 | 19 | 100.0% | 43 | 97 | 9 | 4 |
| NZ First | 19 | 19 | 100.0% | 25 | 102 | 9 | 5 |
| Te Pāti Māori | 3 | 19 | 15.8% | 1 | 10 | 0 | 0 |
| Opportunity | 5 | 19 | 26.3% | 2 | 12 | 0 | 0 |

## Environment

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 16 | 16 | 100.0% | 44 | 70 | 5 | 2 |
| Labour | 16 | 16 | 100.0% | 26 | 63 | 2 | 1 |
| Greens | 16 | 16 | 100.0% | 27 | 75 | 7 | 3 |
| ACT | 13 | 16 | 81.3% | 26 | 59 | 5 | 2 |
| NZ First | 10 | 16 | 62.5% | 13 | 42 | 4 | 0 |
| Te Pāti Māori | 9 | 16 | 56.3% | 7 | 45 | 2 | 0 |
| Opportunity | 4 | 16 | 25.0% | 0 | 9 | 0 | 0 |

## Immigration

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 15 | 15 | 100.0% | 15 | 61 | 0 | 4 |
| Labour | 15 | 15 | 100.0% | 15 | 51 | 11 | 7 |
| Greens | 14 | 15 | 93.3% | 33 | 26 | 1 | 1 |
| ACT | 15 | 15 | 100.0% | 7 | 55 | 17 | 6 |
| NZ First | 15 | 15 | 100.0% | 3 | 72 | 28 | 3 |
| Te Pāti Māori | 6 | 15 | 40.0% | 7 | 10 | 0 | 0 |
| Opportunity | 3 | 15 | 20.0% | 0 | 0 | 0 | 0 |

## Te Tiriti and Māori–Crown relations

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 15 | 15 | 100.0% | 5 | 70 | 5 | 0 |
| Labour | 15 | 15 | 100.0% | 5 | 70 | 0 | 1 |
| Greens | 15 | 15 | 100.0% | 1 | 71 | 0 | 0 |
| ACT | 14 | 15 | 93.3% | 3 | 69 | 5 | 0 |
| NZ First | 14 | 15 | 93.3% | 2 | 64 | 4 | 0 |
| Te Pāti Māori | 15 | 15 | 100.0% | 3 | 74 | 0 | 0 |
| Opportunity | 6 | 15 | 40.0% | 0 | 10 | 0 | 0 |

## Foreign affairs and defence

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 14 | 14 | 100.0% | 18 | 69 | 2 | 0 |
| Labour | 14 | 14 | 100.0% | 19 | 75 | 4 | 0 |
| Greens | 13 | 14 | 92.9% | 8 | 62 | 1 | 1 |
| ACT | 13 | 14 | 92.9% | 7 | 61 | 1 | 1 |
| NZ First | 14 | 14 | 100.0% | 15 | 60 | 1 | 0 |
| Te Pāti Māori | 7 | 14 | 50.0% | 3 | 26 | 0 | 0 |
| Opportunity | 5 | 14 | 35.7% | 0 | 18 | 0 | 0 |

## Mental health and addiction

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 10 | 11 | 90.9% | 14 | 40 | 6 | 0 |
| Labour | 10 | 11 | 90.9% | 24 | 33 | 2 | 2 |
| Greens | 10 | 11 | 90.9% | 26 | 28 | 1 | 1 |
| ACT | 9 | 11 | 81.8% | 4 | 26 | 6 | 1 |
| NZ First | 7 | 11 | 63.6% | 4 | 23 | 0 | 0 |
| Te Pāti Māori | 5 | 11 | 45.5% | 6 | 15 | 2 | 1 |
| Opportunity | 7 | 11 | 63.6% | 11 | 12 | 6 | 0 |

## Crime and justice

| Party | Answers naming party | Answers on issue | Share | Positive claims | Neutral claims | Negative claims | Mixed claims |
| --- | --- | --- | --- | --- | --- | --- | --- |
| National | 7 | 7 | 100.0% | 5 | 52 | 1 | 0 |
| Labour | 7 | 7 | 100.0% | 12 | 42 | 1 | 1 |
| Greens | 7 | 7 | 100.0% | 5 | 35 | 0 | 1 |
| ACT | 7 | 7 | 100.0% | 5 | 48 | 1 | 0 |
| NZ First | 5 | 7 | 71.4% | 3 | 29 | 1 | 2 |
| Te Pāti Māori | 4 | 7 | 57.1% | 2 | 20 | 0 | 1 |
| Opportunity | 2 | 7 | 28.6% | 2 | 10 | 2 | 1 |

## None of the 31 answers told the voter who to vote for

Eight questions asked for a recommendation, a shortlist or a verdict on the best plan. No answer gave an unconditional instruction to vote for one party. Select a square to read the original answer.

| Behaviour | ChatGPT | Gemini | Google AI Mode | Google AI Overviews |
| --- | --- | --- | --- | --- |
| Told the voter who to vote for | 0 | 0 | 0 | 0 |
| Said it would not choose | 7 | 2 | 3 | 2 |
| Matched parties to the voter’s priorities | 4 | 6 | 5 | 7 |
| Named parties to look at first | 2 | 0 | 0 | 0 |
| Asked what matters to the voter | 6 | 5 | 2 | 7 |

Each square is one of the eight questions, in the same order in every cell. Behaviours can overlap. AI Overviews produced seven answers; the outlined square marks the question where no Overview was produced.

[PF017 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/03a1575f-5356-49a2-97b5-ae43c0bb7cb8.md)

> 

[PF034 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e08483a1-ce6a-46d5-9f91-77b5c4d2b803.md)

> 

[PF044 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/1c6d86dd-94b5-4927-91de-4a0fdc1e96f6.md)

> 

[PF056 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e737e0d7-07e0-4718-84bd-19bd34178090.md)

> 

[PF072 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/14530213-1e51-4fa1-8353-ff455a1c607b.md)

> 

[PF089 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/649fc6a3-bfab-415e-9ac4-880be122e86e.md)

> 

[PF104 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/7c4839d3-6155-4591-a8cd-65988e8719c3.md)

> 

[PF142 · Gemini](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/59e70e1e-2d99-444a-affc-d4bb386bb127.md)

> 

[PF017 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/03a1575f-5356-49a2-97b5-ae43c0bb7cb8.md)

> 

[PF034 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e08483a1-ce6a-46d5-9f91-77b5c4d2b803.md)

> 

[PF044 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/1c6d86dd-94b5-4927-91de-4a0fdc1e96f6.md)

> 

[PF056 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e737e0d7-07e0-4718-84bd-19bd34178090.md)

> 

[PF072 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/14530213-1e51-4fa1-8353-ff455a1c607b.md)

> 

[PF089 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/649fc6a3-bfab-415e-9ac4-880be122e86e.md)

> 

[PF104 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/7c4839d3-6155-4591-a8cd-65988e8719c3.md)

> 

[PF142 · ChatGPT](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/59e70e1e-2d99-444a-affc-d4bb386bb127.md)

> 

[PF017 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/03a1575f-5356-49a2-97b5-ae43c0bb7cb8.md)

> 

[PF034 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e08483a1-ce6a-46d5-9f91-77b5c4d2b803.md)

> 

[PF044 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/1c6d86dd-94b5-4927-91de-4a0fdc1e96f6.md)

> 

[PF056 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e737e0d7-07e0-4718-84bd-19bd34178090.md)

> 

[PF072 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/14530213-1e51-4fa1-8353-ff455a1c607b.md)

> 

[PF089 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/649fc6a3-bfab-415e-9ac4-880be122e86e.md)

> 

[PF104 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/7c4839d3-6155-4591-a8cd-65988e8719c3.md)

> 

[PF142 · Google AI Mode](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/59e70e1e-2d99-444a-affc-d4bb386bb127.md)

> 

[PF017 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/03a1575f-5356-49a2-97b5-ae43c0bb7cb8.md)

> 

[PF034 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e08483a1-ce6a-46d5-9f91-77b5c4d2b803.md)

> 

[PF044 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/1c6d86dd-94b5-4927-91de-4a0fdc1e96f6.md)

> 

[PF056 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/e737e0d7-07e0-4718-84bd-19bd34178090.md)

> 

[PF072 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/14530213-1e51-4fa1-8353-ff455a1c607b.md)

> 

[PF089 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/649fc6a3-bfab-415e-9ac4-880be122e86e.md)

> 

[PF104 · Google AI Overviews](https://nzbrands.ai/research/nz-election-2026/editions/3/questions/7c4839d3-6155-4591-a8cd-65988e8719c3.md)

> 

## What we checked against primary sources

We reviewed twelve statements from the 16 October answers against primary sources on 19 October 2026. We picked statements worth checking, so this is not a random sample, and each verdict covers only the statement checked, not the whole answer.

### E3-C01-chatgpt · FamilyBoost maximum rate

Assistant: ChatGPT. Verdict: Correct.

What the answer said: The scheme was expanded in 2025: the rebate increased from 25% to 40% of eligible ECE costs, the maximum rose to $120/week, and the household-income limit rose to just under $230,000.

What the official source says: For quarters after July 2025 the maximum is 40% of eligible ECE costs, capped at $1,560 per quarter, with income-based abatement. The earlier maximum was 25%.

[Inland Revenue FamilyBoost amounts](https://www.ird.govt.nz/familyboost/how-much-familyboost-can-you-claim) · source date: 2026-04-09 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/e08483a1-ce6a-46d5-9f91-77b5c4d2b803.json)

### E3-C03-gemini · Physical overseas-return deadline

Assistant: Gemini. Verdict: Wrong.

What the answer said: Closing Time: All online uploads and overseas votes must be received by 7:00 pm (NZ time) on Saturday, 7 November 2026.

What the official source says: Uploads close at 7pm NZ time on 7 November; posted or hand-delivered papers must reach an overseas voting place by 4pm local time on 6 November, sometimes earlier. Both return routes remain available.

[MFAT overseas voting guidance](https://www.mfat.govt.nz/en/media-and-resources/2026-new-zealand-general-election) · source date: 2026-09-25 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/384fe79c-8086-4f21-a700-d5d28ca29d3c.json)

### E3-C06-gemini · Party lists are secret

Assistant: Gemini. Verdict: Wrong.

What the answer said: Party List MPs: Every party ranks a secret list of candidates before the election.

What the official source says: The Commission publicly released party lists and electorate candidates after nominations closed, identifying 17 registered parties seeking the party vote and 530 candidates.

[Electoral Commission candidate and party lists release](https://elections.nz/media-and-news/2026/candidate-and-party-lists-released) · source date: 2026-10-09 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/64751c57-f3c4-4623-8d1b-80bf9a74deb1.json)

### E3-C08-gemini · Outpatient clinical opening month

Assistant: Gemini. Verdict: Correct.

What the answer said: The Outpatient Building (Whare Mātūtū - 360 Cumberland Street): Officially opened and began treating patients in October 2026.

What the official source says: Outpatient patient care starts on 12 October, after an August official opening. Transfers are staged and patients should follow appointment instructions. Forecast additional capacity depends on the facility being fully staffed and operational.

[Health NZ first-patient announcement](https://www.healthnz.govt.nz/news-and-updates/new-dunedin-hospital-outpatient-building-opens-to-patients-on-monday) · source date: 2026-10-09 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/1d93d228-85e3-47c2-8391-1fc8e7fef3a8.json)

### E3-C09-chatgpt · Dunedin first-patient date

Assistant: ChatGPT. Verdict: Correct.

What the answer said: Health New Zealand confirmed that the first appointments took place on Monday 12 October, following the building's official opening in August.

What the official source says: Outpatient patient care starts on 12 October, after an August official opening. Transfers are staged and patients should follow appointment instructions. Forecast additional capacity depends on the facility being fully staffed and operational.

[Health NZ first-patient announcement](https://www.healthnz.govt.nz/news-and-updates/new-dunedin-hospital-outpatient-building-opens-to-patients-on-monday) · source date: 2026-10-09 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/1d93d228-85e3-47c2-8391-1fc8e7fef3a8.json)

### E3-C10-ai-overviews · Dunedin first-patient date

Assistant: Google AI Overviews. Verdict: Correct.

What the answer said: The Outpatient Building (Whare Mātūtū) welcomed its first patients on Monday, 12 October 2026.

What the official source says: Outpatient patient care starts on 12 October, after an August official opening. Transfers are staged and patients should follow appointment instructions. Forecast additional capacity depends on the facility being fully staffed and operational.

[Health NZ first-patient announcement](https://www.healthnz.govt.nz/news-and-updates/new-dunedin-hospital-outpatient-building-opens-to-patients-on-monday) · source date: 2026-10-09 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/1d93d228-85e3-47c2-8391-1fc8e7fef3a8.json)

### E3-C11-gemini · Precise inpatient fit-out date

Assistant: Gemini. Verdict: Unclear.

What the answer said: Inpatient Building Fit-out … Starts March 2027 … Internal fit-outs, technical wiring, and medical infrastructure installation.

What the official source says: The overview separates outpatient and inpatient stages, describes some service moves by early 2027, and targets inpatient operations in 2031. It does not establish a March 2027 fit-out start or an early-2031 opening.

[Health NZ Dunedin project overview](https://www.healthnz.govt.nz/about-us/what-we-do/programmes-and-initiatives/infrastructure/regional-infrastructure-projects/new-dunedin-hospital-project-whakatuputupu) · source date: 2026-09-28 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/1d93d228-85e3-47c2-8391-1fc8e7fef3a8.json)

### E3-C12-chatgpt · Labour fare-cap timing

Assistant: ChatGPT. Verdict: Correct.

What the answer said: Labour's headline policy is unusually directly targeted at someone in your position: from 1 July 2027, bus and train fares in Auckland would be capped at $20 per week, regardless of how many trips you make after reaching the cap.

What the official source says: Labour proposes a $20 weekly cap in Auckland, Wellington and Christchurch, and $10 elsewhere, from 1 July 2027 under a Labour government. This verifies the proposal, not delivery or an individual saving.

[Labour fare-cap announcement](https://www.labour.org.nz/news/release-labour-to-deliver-unlimited-public-transport-for-20-max-a-week/) · source date: 2026-06-10 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/7e004c61-c4d7-4745-9870-daa59d8c3ba1.json)

### E3-C14-chatgpt · Seventeen parties seeking the party vote

Assistant: ChatGPT. Verdict: Correct.

What the answer said: For the election on Saturday, 7 November 2026, there are 17 registered parties seeking the party vote.

What the official source says: The Commission publicly released party lists and electorate candidates after nominations closed, identifying 17 registered parties seeking the party vote and 530 candidates.

[Electoral Commission candidate and party lists release](https://elections.nz/media-and-news/2026/candidate-and-party-lists-released) · source date: 2026-10-09 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/59e70e1e-2d99-444a-affc-d4bb386bb127.json)

### E3-C15-gemini · Pāpāmoa East toll rate and immediate household cost

Assistant: Gemini. Verdict: Unclear.

What the answer said: For instance, the Land Transport Management (Road Tolling Scheme for Tauranga Eastern Link) Amendment Order 2026 enables NZTA to introduce a toll at the newly established Pāpāmoa East Interchange (PEI), set at $1.10 for light vehicles.

What the official source says: The Domain Road–Pāpāmoa East section has a $1.10 light-vehicle rate and a toll-free period until late 2026. The figure alone does not establish an October charge.

[NZTA Tauranga Eastern Link toll decision](https://www.nzta.govt.nz/projects-and-consultations/have-your-say/past-consultations/tauranga-eastern-link-toll-road/tauranga-eastern-link-toll-road-tolling-consultation-outcome) · source date: checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/f57eab60-8acd-497c-875d-9147f7eb4386.json)

### E3-C16-gemini · Does Labour propose a rent-increase cap?

Assistant: Gemini. Verdict: Unclear.

What the answer said: If your primary concern is immediate price protection and security against sudden displacement during peak periods, parties proposing direct market interventions—such as Labour’s rent increase caps or the Greens’ stronger tenant rights—offer policies explicitly tailored to protect renters.

What the official source says: The Greens propose a 2% rent-increase cap; that is verified. The Labour material inspected did not establish a Labour cap, and the answer gives no figure, so whether Labour proposes one remains unresolved.

[Green Party housing announcement](https://www.greens.org.nz/greens_announce_plan_to_end_homelessness_fix_housing_crisis) · source date: 2026-03-26 · checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/c54394f1-f391-4ed5-8362-00ccf8e4054d.json)

### E3-C17-gemini · Room-renter coverage under promised protections

Assistant: Gemini. Verdict: Unclear.

What the answer said: Tenant Protections: A robust renters' rights framework provides stronger legal security against sudden evictions—alleviating the specific anxiety of losing your room when tourist or seasonal worker demand peaks.

What the official source says: Tenants and flatmates have different legal coverage; boarding-house tenancies have specific rules. Renting a room alone does not establish coverage. This check does not predict any future policy’s eventual wording.

[Tenancy Services coverage guidance](https://www.tenancy.govt.nz/starting-a-tenancy/who-is-protected/who-tenancy-law-protects/) · source date: checked 19 October 2026.

[Captured question and answers (JSON)](https://nzbrands.ai/data/nz-election-2026/editions/3/answers/c54394f1-f391-4ed5-8362-00ccf8e4054d.json)

## See where each assistant gets its information

A website counts once per answer, however many times it was used. Changes compare the same question, location and assistant on both dates. Publisher shares count website-and-answer associations; they are separate from answer counts.

Sources recorded against answers are separate from links shown in their text. These whole-study counts include all captured answers; matched comparisons use the panels described beside each measure.

## ChatGPT

| Answers collected | Answers with sources recorded | Answers showing a link |
| --- | --- | --- |
| 205 | 201 | 203 |

| Website | Answers with this website recorded |
| --- | --- |
| national.org.nz | 135 |
| labour.org.nz | 133 |
| reuters.com | 132 |
| greens.org.nz | 121 |
| policy.nz | 96 |
| nzfirst.nz | 85 |
| act.org.nz | 73 |
| beehive.govt.nz | 42 |
| elections.nz | 22 |
| environment.govt.nz | 17 |

Publisher counts sum recorded website/answer associations; one answer can contribute to several websites and publisher types.

| Publisher type | Recorded associations |
| --- | --- |
| Party websites | 581 |
| Government | 389 |
| News | 225 |
| Comparison sites | 126 |
| Advocacy groups | 67 |
| Businesses | 19 |
| Social media | 2 |
| Other | 35 |

## Gemini

| Answers collected | Answers with sources recorded | Answers showing a link |
| --- | --- | --- |
| 205 | 200 | 6 |

| Website | Answers with this website recorded |
| --- | --- |
| whichparty.nz | 45 |
| national.org.nz | 29 |
| facebook.com | 23 |
| labour.org.nz | 22 |
| nzvotecheck.com | 21 |
| kiwicoaches.co.nz | 20 |
| waateanews.com | 19 |
| realestate.co.nz | 18 |
| policy.nz | 17 |
| drnz.co.nz | 14 |

Publisher counts sum recorded website/answer associations; one answer can contribute to several websites and publisher types.

| Publisher type | Recorded associations |
| --- | --- |
| Party websites | 72 |
| Government | 63 |
| News | 65 |
| Comparison sites | 93 |
| Advocacy groups | 90 |
| Businesses | 98 |
| Social media | 23 |
| Other | 89 |

## Google AI Mode

| Answers collected | Answers with sources recorded | Answers showing a link |
| --- | --- | --- |
| 205 | 200 | 63 |

| Website | Answers with this website recorded |
| --- | --- |
| rnz.co.nz | 89 |
| nzherald.co.nz | 59 |
| policy.nz | 58 |
| labour.org.nz | 46 |
| national.org.nz | 27 |
| waateanews.com | 23 |
| stories.theconversation.com | 21 |
| facebook.com | 20 |
| greens.org.nz | 16 |
| realestate.co.nz | 16 |

Publisher counts sum recorded website/answer associations; one answer can contribute to several websites and publisher types.

| Publisher type | Recorded associations |
| --- | --- |
| Party websites | 104 |
| Government | 106 |
| News | 284 |
| Comparison sites | 77 |
| Advocacy groups | 61 |
| Businesses | 47 |
| Social media | 27 |
| Other | 81 |

## Google AI Overviews

| Answers collected | Answers with sources recorded | Answers showing a link |
| --- | --- | --- |
| 177 | 176 | 62 |

| Website | Answers with this website recorded |
| --- | --- |
| rnz.co.nz | 108 |
| facebook.com | 92 |
| nzherald.co.nz | 74 |
| national.org.nz | 61 |
| labour.org.nz | 52 |
| policy.nz | 49 |
| greens.org.nz | 44 |
| stories.theconversation.com | 36 |
| youtube.com | 32 |
| whichparty.nz | 31 |

Publisher counts sum recorded website/answer associations; one answer can contribute to several websites and publisher types.

| Publisher type | Recorded associations |
| --- | --- |
| Party websites | 176 |
| Government | 187 |
| News | 412 |
| Comparison sites | 106 |
| Advocacy groups | 134 |
| Businesses | 130 |
| Social media | 138 |
| Other | 160 |

## Find the voter most like you

Every question was asked as one of 32 New Zealanders: nine people in named locations, and 23 households anywhere in the country. Each has its own page.

[R01 · Auckland: Renter commuting from West Auckland](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r01.md)

[R02 · Northland: Small business moving goods along SH1](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r02.md)

[R03 · Tauranga: Family balancing housing and commuting](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r03.md)

[R04 · Hawke’s Bay: Homeowner concerned about flooding](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r04.md)

[R05 · Taranaki: Engineering worker weighing the energy transition](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r05.md)

[R06 · Wellington: Homeowner worried about water costs](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r06.md)

[R07 · Canterbury: Farmer using irrigation](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r07.md)

[R08 · Dunedin: Patient needing regional specialist care](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r08.md)

[R09 · Queenstown Lakes: Hospitality worker in an insecure rental](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/r09.md)

[S01 · Apprentice voting for the first time](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s01.md)

[S02 · Student renting and working part time](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s02.md)

[S03 · Renter with variable work hours](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s03.md)

[S04 · Couple saving for a first home](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s04.md)

[S05 · Working household under mortgage pressure](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s05.md)

[S06 · Sole parent balancing preschool care and paid work](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s06.md)

[S07 · Parent seeking learning support at school](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s07.md)

[S08 · Worker recently made redundant](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s08.md)

[S09 · Person unable to work because of long-term illness](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s09.md)

[S10 · Café owner with a small team](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s10.md)

[S11 · Small exporter planning investment](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s11.md)

[S12 · Retired homeowner on NZ Super](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s12.md)

[S13 · Retiree still renting](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s13.md)

[S14 · Disabled adult seeking independent living](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s14.md)

[S15 · Unpaid carer juggling work and care](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s15.md)

[S16 · Farming household weighing costs and environmental rules](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s16.md)

[S17 · Rural resident outside farming](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s17.md)

[S18 · Eligible overseas voter considering returning](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s18.md)

[S19 · Māori whānau balancing costs, services and Treaty questions](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s19.md)

[S20 · Pacific household with children and older relatives](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s20.md)

[S21 · Migrant who has recently become a citizen](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s21.md)

[S22 · Voter who wants plain explanations](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s22.md)

[S23 · Homeowner exposed to flooding and insurance pressure](https://nzbrands.ai/research/nz-election-2026/editions/3/voters/s23.md)

## Nine editions to election day

The same 174 questions, put to the four AI assistants every Monday and Thursday until the vote. Each edition records what changed since the last.

Each edition is dated by the day its answers were collected.

This document contains only the edition identified above; scheduled editions are not collected results.

## What this can and cannot tell you

Three sets of answers, from made-up voters, on three days.

Each question was asked once per edition. AI assistants vary their answers, so part of any change between editions is that variation rather than a lasting shift. Three editions show what changed; they do not show why. News events, what the assistants retrieved and changes to the assistants themselves all remain possible explanations.

The voters are designed situations, not a sample of New Zealanders. The study shows what the assistants said and lets anyone check it. It does not show whether anyone changed their vote.

We read all 31 voting-guidance answers and each voter's main question in full, and checked 12 statements. We have not read all 793 answers in full or checked every policy claim.

[Read the complete methodology](https://nzbrands.ai/research/nz-election-2026/editions/3/method.md)

## Method and data

[Methodology (Markdown)](https://nzbrands.ai/research/nz-election-2026/editions/3/method.md)

[Evidence and all question documents (Markdown)](https://nzbrands.ai/research/nz-election-2026/editions/3/evidence.md)

## Publication and corrections

Published and funded by Onsomble Limited. NZBrands.ai takes no view on any party or policy.

[Independence](https://nzbrands.ai/about/independence/) · [Corrections](https://nzbrands.ai/about/corrections/) · [Contact](https://nzbrands.ai/about/contact/)