Voluntary AI Safety Standard
The government's September 2024 standard of ten voluntary guardrails for AI, superseded on 21 October 2025 by the National AI Centre's Guidance for AI Adoption. Kept here because people still search for it.
- Status
- Voluntary · Australia (Commonwealth) · DISR
- Applies to
- Nobody, as of October 2025. Read the Guidance for AI Adoption instead.
- Key date
- 21 October 2025: Superseded by the Guidance for AI Adoption
- Primary source
- Official document →
- Last reviewed
- 28 September 2026
What it is
The Voluntary AI Safety Standard was published by the Department of Industry, Science and Resources in September 2024: ten guardrails organisations were encouraged to follow when developing or deploying AI, covering accountability, risk management, data governance, testing, human oversight, transparency to users, a way for people to challenge outcomes, supply chain transparency, record keeping and stakeholder engagement. It was never binding. On 21 October 2025 the National AI Centre replaced it with the Guidance for AI Adoption, which the department describes as “updated and simplified guidance that evolves the Voluntary AI Safety Standard”. The standard is history; its ideas survived.
What it requires
Nothing now. While it stood, the ten guardrails asked an organisation to name who is accountable for AI, assess and manage risks, govern its data, test and monitor systems, keep a person able to intervene, tell users when AI is involved, give affected people a way to contest outcomes, know its AI supply chain, keep records, and engage the people affected. The Guidance for AI Adoption folds those ten into six practices without dropping any of the substance: the department calls the change a consolidation, not a reduction in expectation.
Does this reach your business?
Not any more. If a contract, tender or policy you wrote in 2024 or 2025 references the Voluntary AI Safety Standard, update the reference to the Guidance for AI Adoption; the mapping is direct. If you built a governance approach on the ten guardrails, you have already done the work the six practices ask for.
What we recommend
Our advice is to stop citing this standard and cite the Guidance for AI Adoption instead, which is where regulators now look for what “reasonable” AI governance means in Australia. Nothing you did under the ten guardrails is wasted; rename the headings and carry on.
Questions people ask
No. The National AI Centre replaced it on 21 October 2025 with the Guidance for AI Adoption.
No. It was voluntary from the start. The mandatory guardrails proposed alongside it in 2024 were dropped in December 2025 in favour of applying existing law.
They were consolidated into the six essential practices of the Guidance for AI Adoption. The expectations were simplified, not reduced.