DATASETS FRONTIER
How we use AI
This page describes how artificial intelligence features on this site work, what data they can see, and the oversight around them. It starts on 7 November 2026.
What the AI features do
AI features help draft, summarise, compare and review text: answering questions about source material, organising evidence, and preparing drafts for human review. They assist people; they do not certify, approve or decide anything about a person.
Which models and providers
The coach features use Claude models by Anthropic, running through the site owner's own account on his systems. The retention periods and the providers who process data, with their countries, are on the Privacy page.
What the AI can and cannot see
When you use a chat or coach feature, the conversation and the files you select for it are sent to the model provider so it can answer. The enquiry form file is not readable by any AI feature. Nothing you send the AI features is used by us to train models.
Human review stays in charge
Every AI-assisted output is a draft. A person reviews it before it is relied on, and the service is designed to mark results it could not verify instead of filling gaps. See our Source verification approach.
Guidance we draw on
We draw on Australia's AI Ethics Principles and the Australian Government's AI guidance. In practice that means: human oversight of outputs, transparency about where AI is used, respect for privacy, and no AI features that discriminate or cause harm by design.
Limits
AI text can be wrong, incomplete or out of date. Where the site quotes a regulatory source, the verified quote stands over any AI summary. Email [email protected] if you find an output that misreads a source.