Mandatory guardrails for high-risk AI (proposals paper)
A government proposals paper on mandatory requirements for high-risk AI settings. Consultation stage. Nothing in it binds anyone yet.
- Status
- Proposed · Australia (Commonwealth) · DISR
- Applies to
- Would reach developers and deployers of high-risk AI if legislated
- Primary source
- Official document →
- Last reviewed
- 17 July 2026
Certrak’s verdict
The short version
One or two plain sentences: what this is, and whether it bites for a typical business.
What it actually requires
The substance in plain language. Quote the source where the exact wording matters.
Does this reach your business?
Who it applies to, and the common situations where it does or does not.
What sensible looks like
Practical, proportionate steps. Not legal advice: what a reasonable operator would do.