New Zealand hasn’t had a headline-grabbing AI scandal.
That’s Peter Griffin’s opening in a BusinessDesk column from May, and it’s worth sitting with. No deepfake crisis, he writes, no robo-lending meltdown, no self-driving cars going rogue. “Yet, for a country that’s using AI more and more, we don’t really trust it.”
His column is about the AI Forum’s refreshed AI Blueprint for Aotearoa, and it’s behind a paywall, so I’ll stick to what’s free to read. But his framing is the useful bit: “Our trust problem is more about perception than domestic misuse. But perception has consequences.”
I think that’s right, and I think it has a practical consequence for anyone leading a team through this.
What the numbers say
The best public measure we have is the University of Melbourne and KPMG global study, which surveyed about 1,000 New Zealanders among 48,000 people in 47 countries.
Only 34% of New Zealanders said they were willing to trust AI. Just 15% said they highly or completely accept it, the lowest of all 47 countries in the study.
76% said they were concerned about negative outcomes from AI. Far fewer, 33%, said they had personally experienced or observed one.
That’s not nothing. A third of people having seen something go wrong is real, and I’m not going to wave it away. But the concern is more than twice as common as the experience. A lot of the worry is coming from somewhere other than what people have lived through.
The AI Forum’s Blueprint puts it plainly. Across the 28 reports it reviewed, adoption estimates range from 40 to 80 percent, but “trust has not kept pace with uptake.” It calls our position “high-use, low-trust”, and it has made social licence one of two new focus areas.
Where the worry comes from
If it isn’t mostly coming from experience, it’s coming from the noise: headlines from overseas, the friend whose cousin lost a job, the confident person on a podcast, the vague sense that something big is happening and nobody’s saying what.
None of that is irrational. If you can’t see how something is being used around you, imagining the worst is a sensible thing to do. It’s what careful people do.
But it does change what helps.
Why another rule doesn’t settle it
The instinct, when people are uneasy, is to add controls. A new policy. An approval step. A list of things you mustn’t do.
Some of that is worth having. But it doesn’t touch fear that came from outside the building. People have never seen the thing the rule is protecting them from, so the rule doesn’t reassure them. Sometimes it does the opposite: if the business needs this many rules, there must be something to be scared of.
What reaches that kind of worry is something real to measure it against, and that comes from how leaders behave, week after week.
What steady looks like
None of this is clever. It’s mostly consistency:
- Say what AI is for here, and what it isn’t for. In plain words, in a staff meeting, not only in a document. People fill silence with the worst version.
- Do it the same way in every team. Nothing feeds unease faster than one manager encouraging AI use and another quietly frowning on it.
- Tell people about changes before they hear it elsewhere. If a role or a process is going to change, the person in it should hear it from you, early, with time to ask questions.
- Let people see it working. A colleague showing how they used AI on a real task does more than any reassurance from the top.
- Ask. If nothing has gone wrong at work and people are still uneasy, find out where it’s coming from. You may be surprised.
The caveats
The KPMG study is a public survey, not a workplace one, and the fieldwork ran from late 2024 to early 2025, so attitudes may have moved since. The 33% includes things people have “observed”, which could be anything from a bad chatbot to a news story. And Griffin’s column is commentary; I’ve only used the part outside the paywall.
Declaring an interest
Helping businesses bring AI in so their people feel accompanied rather than watched is what we do at Coreshift, so weigh my conclusion with that in mind.
But you don’t need anyone’s help to do the most important part. When the worry is ahead of the evidence, the most reassuring thing in the room is a leader who does the same thing, the same way, every time, and says so.