Election 2026: What parties say about the threat of AI to fool voters

New Zealand's major political parties are drawing a rare line of agreement ahead of this year's election, saying artificial intelligence must not be deployed to deceive voters – although their approaches differ on how the technology will be used during the 2026 campaign.

Some parties are ruling out generative AI entirely; others support limited or clearly labelled use; and some see value in using AI to support campaign work as long as humans stay in control. One party says its MPs have already been the target of AI‑fabricated content designed to stoke racial tension.

Here's each parties position on AI use in the 2026 election campaign below:

National said AI generated content might appear in its social media posts.

"From time-to-time, National uses AI-generated content in its social media. We strive to ensure this content is marked as such," a party spokesperson told 1News.

During the 2023 election, National admitted using artificial intelligence to generate people in attack advertisements on the then Labour government.

The party said it did not condone using AI to depict real people in misleading ways.

"Unfortunately, we do see this type of content from others and expect we will see it during the election campaign," the spokesperson added.

"We urge people to be mindful that not all content they see is real and to check their facts thoroughly before they engage with it."

Labour said it would be taking a trust-first approach to artificial intelligence in this year's election.

Campaign manager Beth Houston said the party wanted to ensure voters knew they were seeing genuine material rather than manipulated or synthetic content.

"Our priority is ensuring the public can continue to trust the information they receive during elections. Labour does not use AI to generate images or videos. Our focus is on authentic engagement with voters."

She told 1News that political parties should disclose when AI was used to generate images and video.

"Misinformation of any sort is always a concern. Using AI to generate political messaging designed to mislead voters would undermine trust and transparency."

ACT said it used artificial intelligence to assist with drafting, content creation and policy research.

"Any AI-assisted material is reviewed closely, fact-checked, and edited by our team before being released for the wider public," said campaign manager Callum Purves.

He said ACT would expect political parties to disclose when AI has been used to produce campaign material if it had the potential to mislead.

"As a rule, ACT does not use AI to generate false images of real people. And, while we do use AI as a tool to supplement our library of photos and stock images, our preference is to use real photography where possible. We believe New Zealand's challenges are serious enough to warrant the human touch."

He added that ACT was concerned by the potential for AI to mislead voters during an election campaign, saying: "Bad actors who spread AI-assisted misinformation should be held to account by the Electoral Commission and the media."

The Green Party held the strongest anti-AI position, pledging not to use generative AI content in its 2026 campaign.

"We have ethical, privacy, and environmental concerns about the use of generative AI and the tech giants and billionaires who are pushing it… we don't plan on using generative AI to create our campaign materials," a spokesperson told 1News, adding it would rely on human artists and creatives.

The Greens said it was "impossible" to ensure the campaign would not use AI tools but staff would be encouraged to avoid them where possible.

"The Green Party recognises that the ubiquity of AI makes it hard for each use to be disclosed but we would expect parties to be transparent about how they are using AI and not mislead the public," the spokesperson said.

"We also hope all parties will join us in calling out misinformation and deepfakes during the election. It is important that the public can trust in the information that they see."

New Zealand First said it was open to the "limited responsible use" of AI for creative visuals such as memes or images but not in strategy or messaging.

"We will not be using AI for the drafting of any scripts, strategies, slogans or messaging, and certainly not used to create misleading deepfakes of politicians or political parties," campaign manager Darroch Ball told 1News, adding anything created with AI would be publicly labelled as such.

He said New Zealand First was "certainly wary" of the potential of using AI to create misinformation and deepfakes where it was deliberately used to show actions or words that had never been made or spoken.

"The biggest concern in this regard is not from other political parties but thewidespread use online from anyone with the use of AI posting on social media," he said.

"This type of deliberate abuse of AI with the sole intention to create misleading, false images, audio, or clips of any politician is of high concern and I think will be a unique challenge for this election."

Te Pāti Māori says any use of artificial intelligence must be guided by tikanga and data sovereignty.

The party said AI had already been weaponised against its MPs and Māori communities, with a spokesperson telling 1News the party was frequently targeted with fabricated images and posts.

“As an indigenous party, our MPs are constantly the subjects of fake AI images and articles designed to fuel anti Māori sentiment for political gain. This causes direct harm to the communities we represent,” the spokesperson said.

The party accepts AI was now “part of the political landscape” but insists its use must be responsible, transparent, disclosed and must not breach tikanga.

“We do not support the use of AI in spreading misinformation, degrading political discourse, or reproducing images of people, art, and mātauranga Māori. The rate at which AI can produce misinformation also poses a risk to democracy and causes real harm to marginalised communities who are most often targeted by online abuse."

The election will be held on November 7 – with overseas voting scheduled to begin on October 21.