Community Stories & Events

Hundreds of Possible Futures for New Zealand's Extreme Rainfall

Earth Sciences New Zealand researchers used generative AI to extend physics-based climate simulations, revealing a wider range of possible rainfall extremes as warming increases.

Who Decides When New Zealand's Public Systems Use AI?

New Zealand's light-touch AI policy leaves important questions about automated public decisions, accountability and Māori data sovereignty unresolved. Andrew Lensen sets out practical models for stronger safeguards.

How Much Warning Is Worth a False Alarm? Machine Learning and New Zealand's Sudden Eruptions

University of Canterbury and Auckland researchers tested a machine-learning eruption forecaster on real seismic data and found it caught four of five Whakaari eruptions, at the cost of about 15 false-alarm days a year.

Teaching AI to Ignore the Birds: Finding New Zealand's Last Possums by Sound

AUT researchers built a training method that cuts false alarms when AI scans forest audio for pest possums, a bottleneck in the hardest 'mop-up' stage of eradication.

Could AI Data Centres Support New Zealand's Grid Instead of Straining It?

Two energy researchers argue that New Zealand's incoming AI data centres do not have to be a burden on the electricity system. Designed to shift when they draw power, they could help absorb wind and solar and underwrite community renewables.

Where AI Might Actually Help Mental Health Care: Spotting Depression Earlier

University of Auckland researchers are testing whether AI can read speech and text for early signs of depression, positioned as a screening aid for stretched services rather than a replacement for clinicians.

A New Zealand Trial Asks Whether Machine Learning Can Personalise Life Support

A New Zealand-led trial, the REVOLUTION trial, has won a $5 million grant to test whether machine learning can help intensive care doctors decide how much oxygen each patient on life support should receive. It will run across 50 ICUs in New Zealand and Australia and recruit more than 24,000 patients.

Clearing the Camera-Trap Backlog: AI and Australia's Wildlife Data

University of Queensland researchers have built WildObs, a cloud platform that uses AI to read millions of wildlife camera-trap images and turn them into conservation data, ten times faster than people.

Bogdan State brought the view from inside the work: thirty years of writing software, and a year spent building alongside an AI coding assistant. His account was refreshingly grounded, and more useful for it.

Martin Kay has watched the same pattern repeat for decades. A new technology arrives, and suddenly every strategy is rebuilt around it. His message at AI for Good Wellington was a useful corrective: AI is a technology you put to work behind a strategy, not the strategy itself.

Michelle Burke put a question on the table that is hard to set down. Expertise is built by doing imperfect work. If AI now absorbs exactly that layer of work, where does the next generation of experts come from?

KiwiHacks: How a 17-Year-Old Built NZ's First Teen Hackathon Scene

Niko Purdie started coding at 9, competed at a hackathon in Austin, Texas, and came home to realise New Zealand had nothing like it for teens. So he built one. KiwiHacks has grown from 10 attendees to 60+ across four events, with a 91 NPS and a weekly coding club that teaches beginners to ship real software.

Wellington GenAI Group

The Wellington Generative AI Community is a space for enthusiasts and learners to explore, share, and deepen their expertise in generative AI. We welcome anyone curious or passionate about this field.