'Use AI in ways you can clearly explain': key tips on maintaining trust with AI use
Mon, 20th Jul 2026
Artificial intelligence is moving fast. But according to Kate Kolich, the organisations that will earn lasting trust are not the ones moving fastest - they're the ones moving most responsibly.
Speaking at a recent panel on building equity and trust in public services, Kate drew on years of experience across banking, the social sector, and energy to lay out what responsible AI and data use actually looks like in practice. Her message was clear: the foundations matter as much as the technology.
Start With the Guard Rails
Before an organisation reaches for an AI solution, Kate argues it needs to know where it stands. That means having a risk appetite statement - a document that sets out clearly how much risk the organisation is willing to take with data, and at what point decisions require escalated governance.
"Put in place the guard rails so everyone inside your organisation knows the rules of the game."
From there, the architecture builds out: technology policies that align with those boundaries, data and information governance frameworks, and a privacy impact assessment process that goes well beyond checkbox compliance. Critically, Kate urges practitioners to conduct privacy assessments not just from their own institutional perspective, but from the perspective of the people whose data is actually being used.
"Put yourself in the shoes of the people represented in the data."
If there is an opportunity to survey customers or service recipients about how they feel about your use of their data, take it - and treat it as genuine input, not a formality.
AI Needs a Policy - and Everyone Needs to Know It
With AI now embedded across industries, Kate raised a pointed question to the room: how many organisations actually have a responsible AI usage policy in place? The show of hands was modest.
Her advice was unambiguous: implement one, make training on it mandatory for every staff member, and then go a step further - publish your AI principles externally, so the people interacting with your organisation know how their data is being handled.
Transparency isn't just an internal governance matter. It's a public commitment.
Explainability and Transparency Is Non-Negotiable
Of all the principles Kate outlined, perhaps the most direct was this: if you cannot explain what your AI solution is doing, you should not be processing people's data with it.
That explainability, she stressed, cannot live solely with a CTO or technical specialist. The ability to articulate what a system does and why - in plain language - needs to sit with the business owner involved with the solution. Trust is built by people who understand the purpose of a tool, not just its mechanics.
"The last thing you want is to have your CIO as the always go-to person that has to explain how an AI solution is working."
-
Ultimately, Kate's message was not about slowing down AI adoption, but about adopting it with discipline and a robust foundation. As organisations face growing pressure to embrace new technologies, trust will increasingly depend on the decisions made long before a solution is deployed.
We are excited to hear Kate again at the upcoming NZ CIO Summit this August; she will be joining the panel "Data as the Backbone of Sovereign AI and Digital Trust" to discuss how organisations can unlock the value of data while strengthening governance, trust, and readiness for the AI era.
CIO Innovation Summit & Awards
4 - 5 August 2026 | NZICC, Auckland
-
Kate teaches a course on responsible data and AI usage for organisations – learn more.