Australia Sets AI Consumer Safety Priorities
Australia has announced five AI consumer-safety workstreams, including an intention to legislate a Digital Duty of Care. This is a policy announcement, not a new enforceable AI rule, and the release does not set a bill text, start date, penalty, or settled scope.
What happened: On 20 July 2026, Australian ministers outlined a whole-of-government set of AI safety priorities.
What is signalled: The package includes planned Digital Duty of Care legislation, a federal-agency automated-decision framework, workplace work through a tripartite forum, privacy reform consultation and examination of consumer-law options.
What has not happened: The release does not enact a new AI duty. It does not identify covered firms, a commencement date, an enforcement agency, a penalty or final operational requirements.
- Authority
- Australian Government, joint ministerial release.
- Instrument
- “AI Consumer safety priorities” media release.
- Event date
- 20 July 2026.
- Status
- Policy direction and workstreams. Not an enacted AI statute or regulation.
- Primary source
- Australian Government joint release
What did Australia announce?
The Australian Government said it was outlining AI safety priorities intended to strengthen protections, increase accountability and support safe adoption across the community and Commonwealth Government. It is an announcement of direction and ongoing work, not the text of a new law.
One stated priority is to legislate a Digital Duty of Care. The release says the intended aim is to put the onus on AI companies to build safety by design and proactively address potential harm. That wording is important as a policy signal, but it does not answer the legal questions a bill would need to settle. The source does not identify the entities that would be covered, the conduct that would trigger a duty, a regulator, a commencement date, defences, remedies or penalties.
For that reason, a business should not state that it has acquired a new Australian statutory Digital Duty of Care as a result of this release. The legally accurate description is narrower: the government has said it plans to legislate one.
What are the five workstreams?
The release names five linked strands. First, the Digital Duty of Care is the legislative workstream. Second, the Attorney-General is to consult on a second tranche of privacy reform. Third, AI safety in the workplace is one of five agreed priority areas for the tripartite Artificial Intelligence Workplace and Employment Forum. Fourth, the consumer-protection workstream will examine options under Australian consumer law for risks including retail surveillance pricing and agentic commerce. Fifth, the Attorney-General is to develop a framework to better regulate automated decision-making within federal agencies.
Each verb matters. “Legislate,” “consult,” “pursue,” “examine,” and “develop” describe different stages of government activity. None of the latter four terms establishes a final legal requirement. The release supplies no consultation dates, draft instruments or compliance deadlines for those streams.
| Workstream | What the official release supports |
|---|---|
| Digital Duty of Care | The government says it will legislate. No final duty, scope or start date is published. |
| Privacy | A second privacy-reform consultation is planned. |
| Workplace AI | AI safety is a priority in a tripartite employment forum. |
| Consumer law | Government will examine options for risks including surveillance pricing and agentic commerce. |
| Federal automated decisions | Government will develop a framework for federal-agency use. |
What does this mean for companies and public-sector teams?
The immediate task is horizon scanning, not a new compliance filing. Teams with Australian products can preserve a clear internal record of consumer-facing AI uses, safety controls, decision paths and complaints patterns. That information may be useful if a future bill, consultation or consumer-law proposal creates a defined process. It does not become a present legal reporting or audit obligation merely because the policy release describes intended direction.
The federal automated-decision workstream is also limited in the source. It concerns use within federal agencies and is described as a framework to better regulate that use, with fair, accurate and transparent decision-making as an objective. The release does not establish a general private-sector automated-decision law, and it does not specify the future framework's content.
For organisations that work with Commonwealth agencies, the prudent distinction is between a potential future public-sector procurement or governance development and a current mandate. Contracts, guidance and legislation would need to be checked when issued. This release alone does not create a new contract term or technical standard.
How does the AI Safety Institute fit in?
The government says the priorities build on the AI Safety Institute, which it says has commenced frontier-system safety testing and established research partnerships and collaborations. These descriptions report operational and research activity. They do not transform testing or collaboration into a new generally applicable approval regime for AI systems.
The same caution applies to the release's reference to existing criminal offences regarding non-consensual deepfake sexually explicit material and work to ban nudify apps. The release distinguishes those existing or ongoing measures from the forward-looking priority package. A reader should not merge them into a single new AI law.
Australia has signalled a policy path toward stronger AI consumer protections, including planned Digital Duty of Care legislation. The 20 July release is not that legislation. Track later bill text, consultations and agency materials before treating any new obligation as settled.
AI Consumer safety priorities, Australian Government joint media release, 20 July 2026. To verify, confirm the five listed workstreams and the release's future-facing verbs.
FAQ
Has Australia enacted a new AI Digital Duty of Care?
No. The 20 July 2026 release says the government will legislate a Digital Duty of Care. It does not publish enacted legislation, a commencement date, covered entities, penalties or final operational requirements.
What are Australia's five AI consumer-safety workstreams?
The release lists planned Digital Duty of Care legislation, privacy-reform consultation, workplace AI safety work through a tripartite forum, examination of consumer-law options, and a framework for automated decision-making in federal agencies.
Does the announcement create a new private-sector AI audit or reporting duty?
No. The release does not identify a new audit, reporting duty, technical standard, penalty or compliance deadline for private-sector AI systems.
What should an organisation do now?
Treat the announcement as a horizon-scanning signal. Keep an evidence-based record of relevant consumer-facing AI uses and controls, and wait for official bill text, consultations or agency guidance before treating requirements as binding.