AI tool
Any computer software that uses artificial intelligence in its processing.
How the laws define this
Both frameworks establish that an AI tool centers on software or systems performing operations driven by artificial intelligence or exhibiting intelligent behaviour. The primary divergence between them lies in their level of detail, scope, and direct reference to regulatory principles. The Artificial Intelligence Security Policy (DWP) relies on a high-level, simple standard, defining an AI tool straightforwardly as any computer software that uses artificial intelligence in its processing. Conversely, the Office of the Privacy Commissioner — Guidance/Expectations on the use of AI (privacy guidance) sets out a broad working definition structured around distinct practical applications. It explicitly itemizes specific categories, including machine learning models trained on data, classifiers, speech-to-text or image captioning interpreters, text, image, or code generative systems, and automated decision systems. Additionally, the privacy guidance uniquely highlights lifecycle compliance by explicitly noting that these tools potentially engage Information Privacy Principles across their entire lifecycle.
Synthesised from the 2 statutory definitions below. Each one is quoted in full, with its source.
How jurisdictions differ
What is distinctive in each definition, binding law in force first.
| Jurisdiction | Instrument | Status | What is distinctive |
|---|---|---|---|
| United Kingdom | Artificial Intelligence Security Policy (DWP) | In Force | This definition limits AI tools to computer software, excluding hardware or other systems. |
| New Zealand | Artificial Intelligence and the Information Privacy Principles | In Force | This definition expands to include specific AI systems like machine learning models, classifiers, generative systems, and automated decision systems across their lifecycle. |
What it triggers
- In the United Kingdom under the DWP Artificial Intelligence Security Policy, it sets out responsibilities for secure, responsible, and transparent use of AI tools by employees, contractors, and suppliers for official DWP business.source
- In New Zealand under the Artificial Intelligence and the Information Privacy Principles guidance, it calls for privacy impact assessments, human oversight, and engagement with Māori principles when using AI systems involving personal data.source
Definitions (2)
Any computer software that uses artificial intelligence in its processing.
A broad working definition used by the guidance to cover systems that perform tasks that appear to exhibit intelligent behaviour, including machine learning models trained on data, classifiers, interpreters (speech-to-text, image captioning), generative systems (text, image, code) and automated decision systems. The guidance treats these systems as potentially engaging the Information Privacy Principles across their lifecycle.
Questions and answers
What security measures apply to AI tools in UK DWP?
The DWP AI Security Policy outlines responsibilities for using AI tools securely, responsibly, and transparently in official business.
How does New Zealand guidance apply AI to privacy?
The guidance applies the Information Privacy Principles to AI systems throughout their lifecycle, covering various types like generative and decision systems using personal data.
What types of AI are covered in these definitions?
Definition 1 covers any software using AI processing, while definition 2 specifies examples like ML models, classifiers, generative systems, and automated decisions.