New Zealand Police Use Two AI Translation Models to Reduce Errors

Oct 23, 2026
New Zealand Police are using Meta's Seamless and OpenAI's Whisper for internal transcription and translation, allowing investigators to choose the more accurate model for each recording.

New Zealand Police are using Meta's Seamless alongside OpenAI's Whisper for transcription and translation, according to RNZ. Police added Seamless around May, while Whisper has been in use since 2024.

Both models run through an internal platform called TransScriptor. Investigators can select the model that provides the most accurate result for a particular language, dialect, recording quality or use case, reducing reliance on one system.

Users submit audio through a web-based interface, and police say no data is transmitted outside the organisation. The platform records who submitted each request, along with the recording length, language and number of jobs processed. Police plan to add the tools to their Technology Capabilities List this month.

RNZ also reported that police set up an AI lab a year ago to test generative AI in an isolated sandbox using historic data, at a time when officers had been breaching AI safeguards and exposing sensitive information.

We hope you enjoyed this article

Consider subscribing to one of our newsletters like AI Policy Brief or Daily AI Brief.

Also, consider following us on social media:

AI Safety & Regulation AI Brief AI Brief (X)
Free newsletter

AI Policy Brief

Weekly report on AI regulations, safety standards, government policies, and compliance requirements worldwide.

AI Token Webinar

Whitepaper

Tensordyne Napier: What If One Rack Could Do the Work of Nine?

Tensordyne

This Tensordyne whitepaper presents Napier, an inference-focused AI processor and rack-scale system based on the company’s TDN Math logarithmic number system. It examines infrastructure requirements for large mixture-of-experts and agentic models, compares major inference architecture approaches, and details the TDN AIP processor, TDN72 pod, TDN Link fabric, and Napier Ultra configuration. The paper reports simulation-based performance, cost, and accuracy-validation results, including Tensordyne’s projected comparison of one Napier rack with a nine-rack Nvidia Rubin plus Groq deployment; the chip is reported as taped out and in fabrication.

Read more

You may also like

Free, six days a week

Daily AI Brief: the AI news that matters, in your inbox.