AI training at work: who’s responsible, you or your staff?

Ask most business leaders who should be responsible for building AI skills, and you tend to get one of two answers. Either employees should keep their own skills current, or the business should train its people. When it comes to workers, however, they’ve already made up their minds.

Kiwi workers are the most likely of any market in our research to say AI training is the employer’s job, not something they should teach themselves. While workers in other countries lean toward self-taught skills, New Zealanders are looking directly to their leaders. For business leaders,  that isn’t a burden; it’s an opportunity to take the wheel.

What is actually happening right now

While the debate about responsibility plays out, the training is happening (and not where you’d expect it). More than half (51%) of New Zealand workers are teaching themselves AI skills via social media. They’re picking up prompts, tips and tools from whatever surfaces in their feed.

Meanwhile, formal support lags behind. Just 17% of NZ employers consider AI skills as essential, while 33% see them as a bonus. And when hiring, just 24% cite AI skills as entry-level criteria, behind Australia at 30%. The message staff receive is mixed: AI is useful but not quite expected, valued but not quite resourced. Faced with this ambiguity, workers default to a patchwork learning approach.

Why leaving it to staff is riskier than it looks

Self-teaching isn’t a bad instinct, but when unmanaged at work, it can lead to real operational exposure: 

• Inconsistency. Ten employees learning from ten different sources end up with ten different standards. There’s no guarantee that they’ll all match.
• No guardrails. Social media algorithms rarely cover data privacy, client confidentiality or when a human needs to check the output. Yet those are exactly the things a business can’t afford to get wrong.
• Hidden usage. Without clear boundaries, employees operate in the dark. In our research 38% of Kiwi workers admit using AI for parts of their job feels like cheating, and 32% admitted to presenting AI-generated work as their own without disclosing it. 

Leaving training entirely to your team doesn’t save time or money; it simply substitutes active leadership with silent liability.

So who’s responsible? Both, but you lead

The honest answer is that AI capability is a shared responsibility. While individuals have to be willing to learn in New Zealand specifically, employers need to lead. 

There are two reasons for this: first, your people already expect you to. That expectation is an opportunity, because a workforce waiting for direction is far easier to move than one that has to be convinced.

Second, only the business can set the standards that matter. That includes what tools are approved, what data is off limits, when to disclose and what good output looks like. By regulating AI use, you ensure that it’s used safely and effectively.

The upside is that Kiwi workers are unusually ready to be led here. Some 57% believe AI is helping them develop valuable skills, against just 34% who fear it makes them replaceable. That’s the lowest replacement anxiety of any market. You don’t have to build enthusiasm from scratch; you just have to give it direction.

A guide to getting started

Businesses don’t necessarily need a large training budget to begin. Most SMEs can make real progress with a phased approach over a single quarter.

What a good program actually covers

Whatever format you choose, a useful AI training effort should cover each of these points:

• The tools your business has approved and how to access them
• What data must never go into an AI tool, in plain language
• When a human has to review or sign off on AI output
• When and how to be transparent that AI was used
• A few concrete, role-relevant examples of AI doing real work well
• Where to ask questions and who has the answers

Notice how much of that is about clarity, rather than technical skill. The single most valuable thing you can give your team isn’t clever prompts or agent setups. It’s permission and boundaries. Once employees know AI is welcome and where the guardrails are, the guilt disappears. Hidden usage comes into the open, giving you the visibility you need to actually manage it.

The takeaway

The responsibility debate resolves itself quickly once you look at the data. Skills are built together, but in New Zealand the employer has to hold the pen. Your workforce is willing to follow and learn, they need your direction.

Businesses that treat AI training as someone else’s job are paying the cost in inconsistency, risk and lost momentum. The ones that step up in the next quarter will turn a ready workforce into a lasting competitive advantage.

For the full picture on how New Zealand workers and employers are approaching AI, including how we compare with Australia, Canada and the UK, read our research: The AI Paradox at Work

The information in this article is current as at 1 July 2026 and has been prepared by Employment Hero Pty Ltd (ABN 11 160 047 709) and its affiliates (Employment Hero). The views expressed in this article are general information only, are provided in good faith to assist employers and their employees, and should not be relied on as professional advice. Some information is based on data supplied by third parties. While such data is believed to be accurate, it has not been independently verified and no warranties are given that it is complete, accurate, up to date or fit for the purpose for which it is required. Employment Hero does not accept responsibility for any inaccuracy in such data and is not liable for any loss or damages arising directly or indirectly as a result of reliance on, use of or inability to use any information provided in this article. You should undertake your own research and seek professional advice before making any decisions or relying on the information in this article.