Acceptable use of Generative AI
This policy establishes the acceptable use guidelines for generative AI tools by New Zealand Police personnel. It mandates mandatory staff training, prior approval of tools by ICT and governance groups, human-in-the-loop oversight, appropriate labeling of AI-generated outputs, and restrictions on court use without prior approval. Furthermore, the policy outlines audit requirements, vendor compliance rules, and a six-month rolling review frequency.
Summary and content fields are drawn from the document text and are pending editorial review.
Who issued it
Who it governs
New Zealand Police staff, contractors, and vendors utilizing or developing generative AI tools for police operations.
What the document contains
This document addresses the internal governance, procurement, risk assessment, training, auditing, and permitted use cases for generative AI tools by New Zealand Police staff, contractors, and external vendors. It includes specific protocols for transcription and translation tools, urgent operational exemptions, mandatory labeling conventions, and accountability measures.
Technical source code specifications, algorithmic training dataset requirements, and public-facing accountability reporting beyond internal governance structures.
- All staff must complete acceptable use of generative AI training before utilizing any genAI tools.
- Every genAI tool requires formal review by ICT and Technology Assurance, along with approval from the Security and Privacy Reference Group.
- Fully automated use of genAI without a human in the loop is prohibited, and users remain fully accountable for outcomes.
- The use of genAI outputs in any court context is prohibited unless prior approval from the court has been obtained.
- All genAI tools must include audit logs tracking prompts and uploaded files, with periodic audits occurring every six months.
· Table of Contents · Executive Summary · Policy statement and principles · Overview · What · Why · Who · How and When · Authority for urgent use · Use of AI transcription and translation tools · Use of Generative AI tools by vendors · Governance, audit and rolling review · Additional Information · Appendix A: Suggested Generative AI Labels · Appendix B: Principles and Risk Matrix for genAI tools
How it scores
Each score is backed by a quote verified against the document text. Editorial review has not yet confirmed them.
Accountability
4/5Clear responsibility for AI outcomes
The user remains fully responsible and accountable for the outcomes and consequences that result from the use of a genAI tool, including for any Code of Conduct breaches.
Who
screening score · 2026-10-23
Fairness
3/5Non-discrimination and equitable treatment
Generative AI transcription and translation tools should be specifically assessed for their accuracy with Māori and Pacific languages before being utilised on audio primarily in these languages.
Use of AI transcription and translation tools
screening score · 2026-10-23
Human Dignity
1/5Respect for the inherent worth of persons
This policy ensures that appropriate safeguards are in place for Police’s use of genAI, and that use of genAI is lawful, ethical, proportionate, and safe in a New Zealand policing context.
What
screening score · 2026-10-23
Human Oversight
4/5Humans remain in control of consequential decisions
New Zealand Police does not currently permit the use of genAI in fully automated contexts (i.e. without a human-in-the-loop).
Who
screening score · 2026-10-23
Privacy
4/5Protection of personal data and private life
privacy and the use of personal and/or sensitive information
Who
screening score · 2026-10-23
Transparency
4/5Openness about how AI systems work and are used
When using a genAI tool, Police staff should label the outputs as being AI generated, referencing the name of the tool, unless there is a clear operational reason to withhold this information.
Who
screening score · 2026-10-23
Enforcement
4/5Consequences and compliance mechanisms with real force
If there are concerns that a user has used genAI beyond the approval scope, this may be considered a disciplinary matter or referred to the Integrity and Conduct group
Audit
screening score · 2026-10-23
Impact Assessment
4/5Required assessment of an AI system's risks/impacts before or during use
Every genAI tool must be reviewed and approved for use by ICT and Technology Assurance - refer to the policy for Trial or adoption of new policing technology.
How and When
screening score · 2026-10-23