Australia's Voluntary AI Safety Standard is a practice framework of ten guardrails for the safe and responsible use of AI, applicable across the AI supply chain. It was released on 5 September 2024 by the Department of Industry, Science and Resources (DISR) and the National AI Centre (then within CSIRO Data61), alongside a Proposals Paper for Introducing Mandatory Guardrails for AI in High-Risk Settings. The standard is positioned as an on-ramp to anticipated mandatory requirements, and it aligns closely with ISO/IEC 42001 and with Singapore's Model AI Governance Framework for Generative AI (Singapore Model AI Governance Framework for Generative AI (2024)). Consultation on the parallel mandatory framework closed 4 October 2024.
Ten guardrails
The Voluntary AI Safety Standard sets out ten guardrails applicable across the AI supply chain:
- Accountability process (governance, internal capability, compliance strategy)
- Risk-management process
- Data governance and AI-system protection
- Testing and monitoring
- Human control and meaningful oversight
- User disclosure of AI-enabled decisions, interactions, and content
- Contestability — a process for people to challenge AI use or outcomes
- Supply-chain transparency among organisations
- Records and documentation for third-party assessment
- Stakeholder engagement with focus on safety, diversity, inclusion, and fairness
Mandatory-guardrails proposal
The companion Proposals Paper, released for parallel consultation, set out three options for mandatory regulation: a domain-specific approach, a framework approach, and a whole-of-economy AI Act. The government's indicative preference, signalled in its January 2024 Interim Response, was for a framework or hybrid approach rather than a whole-of-economy EU-style AI Act (EU AI Act (Regulation 2024/1689)). Under the mandatory proposal, guardrails 1–9 track the voluntary standard closely, while guardrail 10 is replaced by a conformity assessment requirement (audit plus public certification). The definition of "high-risk settings" was left to be enumerated, with the specific list not yet finalised.
Governance and regulators
DISR leads AI policy, and the National AI Centre (CSIRO Data61) provides practice guidance. Australia takes a multi-regulator approach in which existing bodies enforce within their remits: the Office of the Australian Information Commissioner (OAIC) on privacy, the Australian Competition and Consumer Commission (ACCC) on consumer protection, the eSafety Commissioner, and sectoral regulators including the Australian Prudential Regulation Authority (APRA), the Therapeutic Goods Administration (TGA), and the Australian Securities and Investments Commission (ASIC). As of April 2026, Australia did not have an operational AI Safety Institute, in contrast to the UK, US, Japan, Korea, and the EU (via the AI Office); Singapore is also active in this area but operates under a different model. As of April 2026, no mandatory AI bill had been introduced. The legacy of the Robodebt scheme is described in the secondary sources as shaping the Australian political context for AI governance, a point held at medium confidence.
Provenance and confidence
This page summarises Australia's Voluntary AI Safety Standard and its companion mandatory-guardrails proposals paper from the DISR primary materials and a set of law-firm and vendor analyses. The release date (5 September 2024), the ten-guardrail structure, the parallel mandatory consultation closing 4 October 2024, the indicative preference for a framework or hybrid approach, and the alignment with ISO/IEC 42001 are each supported across multiple sources and held at high confidence. The absence of an introduced mandatory bill and the Robodebt context are held at medium confidence.
Relationships
- supports: Australia VAISS (legislation/framework page) — the companion legislation/framework page
- related: Singapore Model AI Governance Framework for Generative AI (2024) — closest analogue; voluntary, horizontal, practice-focused
- related: Data (Use and Access) Act 2025 — Source Summary — parallel voluntary/sectoral posture with a pending comprehensive framework
- related: EU AI Act (Regulation 2024/1689), South Korea AI Basic Act — Source Summary, Japan AI Promotion Act — Source Summary, Brazil AI Bill (PL 2338/2023) — Source Summary, Canada AIDA (Bill C-27, Part 3) — Source Summary — comparator AI-governance regimes
- related: The Bletchley Declaration (AI Safety Summit, 1–2 November 2023), Frontier AI Safety Commitments (Seoul, 2024), Paris AI Action Summit Declaration (2025) — declarations Australia has signed
- related: AI Safety Cases and Frameworks, AI Sovereignty — the multi-regulator approach as a distinct sovereignty posture
Sources
- DISR Voluntary AI Safety Standard:
https://www.industry.gov.au/publications/voluntary-ai-safety-standard - DISR Proposals Paper: Introducing Mandatory Guardrails for AI in High-Risk Settings (Sept 2024)
- Ashurst, "Australia: New AI safety 'guardrails,' and a targeted approach to high-risk settings"
- Herbert Smith Freehills, "Australia releases new mandatory guardrails and voluntary standards on AI"
- White & Case, "Australia launches new AI guidance"
- Corrs Chambers Westgarth, "Australia releases proposed mandatory guardrails for AI regulation"
- Dentons, "Australian AI Reform – How to address the new AI guardrails"
- Securiti, "An Overview of Australia's Voluntary AI Safety Standard"
- FairNow, "Australia's Voluntary AI Safety Standard: Overview"