The register
Six practices, starting with who answers
The guidance asks an organisation to work through all six practices, beginning with the first steps in each and adding the later ones as its use of AI grows. In its foundations version, the very first step is to make a senior leader the overall owner of AI governance.
Decide who is accountable
Give AI governance an overall owner, give every AI system its own accountable person, and write an AI policy for staff.
Understand impacts and plan accordingly
Work out who a system could affect, and give those people a way to report problems and challenge its decisions.
Measure and manage risks
Screen each use, and flag any that carries unacceptable risk or needs closer governance.
Share essential information
Keep a register of every AI system and how it is used, and make telling people about AI use the normal practice.
Test and monitor
Test a system before it goes live, keep watching it afterwards, and ask a supplier to show proof of testing.
Maintain human control
Keep a person overseeing each system, able to pause it, override it, roll it back or shut it down.
Each practice, with every first step the guidance gives, is set out in the six essential practices, step by step.
Five entries
The five guides
Entry 01
The Guidance for AI Adoption
Its two versions, how it relates to the older Voluntary AI Safety Standard, and what each of the six practices asks first.
Entry 02
Buying an AI product
The National AI Centre’s questions for suppliers, and its checks before you commit.
Entry 03
Claims about AI
What the Australian Consumer Law asks of anything a business says about the AI in its product.
Entry 04
Training or fine-tuning on personal information
The OAIC’s October 2024 guidance for anyone training or adapting a generative AI model.
Entry 05
Financial services and credit licensees
ASIC’s Report 798 of 29 October 2024: its eight findings, the existing obligations it points to, and its eleven questions for licensees.
Beside the guidance
The laws a business already answers to
The National AI Centre is plain that AI does not sit outside ordinary law. In a briefing for directors and managers, it says the laws that already govern how a business operates apply to its AI systems as well, naming privacy, anti-discrimination, workplace and intellectual property law among them, and warning that misleading statements in AI-driven advertising could breach consumer law.
For company directors, the same briefing points to section 180 of the Corporations Act 2001, and says the duty of care and diligence there includes making sure adequate governance systems exist to manage the risks an AI system creates. Section 180(1) reads:
“A director or other officer of a corporation must exercise their powers and discharge their duties with the degree of care and diligence that a reasonable person would exercise if they: (a) were a director or officer of a corporation in the corporation’s circumstances; and (b) occupied the office held by, and had the same responsibilities within the corporation as, the director or officer.”
The briefing also says that anyone unsure of their legal obligations around AI should get legal advice, and that even a good lawyer can help only if someone in the business can explain what the system is for, the data it relies on and how it works in practice.
Dated 15 July 2026
Mandatory standards, announced
On 15 July 2026 the Prime Minister announced that the Government will establish “a set of Australian Standards for AI”. He said they would bring the expectations for large AI data centres announced in March “into one regulatory framework”, “Clear, consistent and mandatory”, and that the Government would aim to bring the legislation to Parliament early next year: early 2027.
Among the matters the standards will set rules for, the speech named large data centres: their location, and their use of power and water. It did not set out what the standards would ask of a business that uses AI, or how they would sit beside the Guidance for AI Adoption.
For now, the six practices are what the National AI Centre offers as a shared baseline for responsible AI use, and they sit on top of laws that already apply in full.