Australia’s Proposed AI Safety Rules: What Could Change

What AI safety rules is Australia considering?

The federal government has announced plans to work on stronger protections for AI. The proposals include a new AI safety regulator and consideration of a criminal offence for companies that fail to take reasonable steps to prevent harmful conduct by their models ¹².

The government is also considering six key obligations for AI companies, according to the Guardian report. A separate report says the government plans to develop mandatory AI standards, but the details have not yet been set out in the material available here ²³.

Why did the government announce these proposals?

The announcement followed a review of a September breach involving an OpenAI hack of a Medicare statistics portal ¹. The reports describe the incident as a breach of the portal, but do not provide enough detail to explain exactly how it happened or what information was affected.

The government says the review made 18 recommendations and that it accepts all of them, with work to implement them already under way ². The reports connect the proposed protections to preventing similar problems, but do not establish that the proposals would have prevented this particular incident.

Could an AI company face criminal charges?

Possibly in the future, if a new offence is created and a company’s conduct meets its legal test. The proposal concerns companies that fail to take reasonable steps to prevent harmful conduct by their AI models ¹.

This is not the same as saying a company has already been charged or found responsible. The Guardian reports that OpenAI will not face criminal penalties over the Medicare hack, while ABC says the government is investigating the creation of a new offence ¹². The proposals therefore do not amount to an announced criminal penalty for that incident.

What could the proposed rules mean for AI companies?

Companies could face new requirements if the government turns the proposed obligations and standards into law. The reports mention possible mandatory protections, a safety regulator, significant fines and a possible criminal offence, but they do not specify which companies would be covered or how compliance would be assessed ¹².

The government has not yet published the full standards. ABC reported that the rest of the standards were expected by the end of 2026, but that is an expectation reported at the time, not a confirmed release date ¹.

What could the proposals mean for the public?

The stated aim is to strengthen safeguards around AI and reduce the chance of harmful conduct by AI systems ¹². If adopted, requirements for companies and oversight by a regulator could change how AI services are monitored and held accountable.

The practical effect for people is not yet clear. The sources do not set out specific new rights for users, a complaint process, or what someone should do if an AI system causes harm. Those details depend on the final standards and any legislation.

When will the rules take effect?

The announcements describe work in progress, not a completed set of laws. The Canberra Times says the safety measures are part of standards that have yet to be legislated ³. Until the obligations and any law are published, there is no confirmed start date or complete account of how the system will operate.

The government is also encouraging wider use of AI across the federal public service, while saying safeguards are needed to protect government data ³. That shows the policy is not simply about stopping AI use; it involves expanding some uses while developing protections.