Corrections says any misuse of technology is taken 'extremely seriously'. Photo / RNZ / Blessen Tom
By Sam Sherwood of RNZ
Corrections staff have been warned about the use of artificial intelligence tools after some staff were found to be using it to draft formal reports.
Corrections said any misuse of technology is taken “extremely seriously”, and that they have made it clear to staff that any use of AI tools outside of their approved use is “unacceptable”.
RNZ understands there have been instances where staff used AI to draft formal reports such as Extended Supervision Order reports.
In response to questions from RNZ, chief probation officer Toni Stewart said Corrections’ use of AI was currently limited to Microsoft Copilot. Other publicly available AI applications are blocked on the Corrections network.
“This ensures AI use at Corrections occurs within an environment where we can manage privacy and security controls.”
Staff use of Copilot was governed by its AI policy, which was in line with guidance from the Government’s chief digital officer.
“The policy is explicit that personal information, including any identifying details, health or medical information, or details relating to people in Corrections’ management, must not be entered into Copilot Chat.”
Stewart said the uptake of Copilot remained “relatively low” with about 30% of Corrections staff engaging with the tool since it was introduced on Corrections devices in November 2025.
“Copilot is intended to be used solely as an assistive tool to create and refine content that does not contain sensitive information. Corrections staff can only access the free Microsoft Copilot Chat feature that is part of our existing Microsoft 365 licence and is a standalone chat function, without integration into our system data.”
Stewart said the policy was clear that Copilot Chat must not be used under any circumstances to draft, structure, analyse, or generate content for reports or assessments that contain personal information. Staff may be subject to auditing, with all prompts searchable and exportable.
“We have recently become aware of a small number of incidents where staff have used Copilot to assist with their work in a way that does not comply with our AI policy and guidance.
“We’ve taken action as soon as we’ve become aware of these instances and made it extremely clear that any use of Copilot outside of its approved use is unacceptable.”
A privacy risk assessment was undertaken where instances of inappropriate use were identified.
“Our leaders, particularly within Community Corrections where staff write a number of reports, are actively working to ensure proper AI use is an ongoing conversation with staff.
“Staff are regularly reminded of the AI policy and other relevant guidance.”
Stewart said Corrections was “actively working” to ensure any ongoing use of AI was “safe, secure and appropriate”.
“Corrections has an AI assurance officer, a function held by the director cybersecurity, who is accountable for guiding safe and secure adoption of AI. This includes external reporting to the Government chief digital officer.
“Corrections participates in the All of Government Community of Practice on AI, managed by the Government chief digital officer. We also have established the AI working group to provide formal governance of AI which includes embedding safe and ethical AI practices across the department and provide consistent advice on its safe use.”
Any misuse of technology was taken “extremely seriously”, Stewart said.
“We are committed to protecting the privacy of the people we work with and maintaining the professional integrity of our assessments, reports, and case documentation.”
As of Friday, no notifications had been made to the Office of the Privacy Commissioner, a Corrections spokesperson confirmed.
“Alongside our existing guidance, our privacy team are working with the relevant work groups to provide further guidance on the use of Copilot in the Community Corrections space. Any information entered into Copilot by Corrections remains within the Corrections’ domain.”
A spokesperson for the Office of the Privacy Commissioner (OPC) said in a statement that the Privacy Act applied to the use of personal information, including through AI tools.
It was the responsibility of agencies to understand the technology they use and to ensure use met privacy requirements, the spokesperson said.
“Corrections has stated that its policy prohibits staff from entering personal information into Copilot Chat or using Copilot to prepare reports or assessments containing personal information.
“If this is correct, then privacy concerns would be limited to any cases in which Corrections staff use Copilot in breach of Corrections policy. Where Copilot is used in a way that breaches Corrections policy, OPC would expect Corrections to take appropriate action to remedy this.”
- RNZ