Australia Plans Strict Rules on How Government AI Can Make Decisions About You

Key points

• Australia's Albanese government is drafting a national plan to govern AI use inside government departments and agencies.
• The rules are expected to cover consumer protections, workplace safety and privacy.
• The plan prioritises fairness, accuracy and transparency in automated government decisions.
• A separate Labor push for "digital duty of care" legislation is running alongside the new plan.

When a government agency decides whether you qualify for a benefit, a visa, or a licence, it may not be a human making that call. It may be an algorithm, a set of computer instructions that weighs your details and spits out an answer. Australia's federal government now wants strict rules over exactly that process.

Senior ministers in the Albanese government have begun drafting a national AI plan, first reported by The Guardian, that would apply to automated decision-making across government departments. Automated decision-making simply means a computer system, not a person, reaching a conclusion that affects someone's life.

The plan is expected to reach beyond government offices. Consumer protections, workplace safety rules and privacy safeguards are all on the table.

The push comes as AI use inside public agencies grows fast, and as Australia sees a construction boom in data centres, the large warehouse-like facilities packed with computer servers that store and process AI workloads.

What does this mean for ordinary people?

If you have ever had a government application rejected and wondered why, these rules are meant to give you a clearer answer. The Albanese government says its framework will prioritise three things: fairness, accuracy and transparency. In plain terms, that means the system should treat people equally, get the facts right, and be able to explain its reasoning.

That last point matters most. Right now, automated systems can deny an application without any human reviewing the logic. The new rules aim to change that.

Running alongside the national plan is a separate Labor proposal for what the government calls a "digital duty of care" law. A duty of care is a legal obligation to avoid causing harm. Applied to digital systems, it would mean companies and agencies that deploy AI tools carry a legal responsibility if those tools hurt people.

Neither piece of legislation is final yet. The rules are still being written.

Watch for these signs that automated government decisions may affect you:

• A rejection letter that offers no specific reason, only a reference number.
• A decision that arrives unusually fast, with no case officer named.
• A request to reapply with more documents, but no explanation of what went wrong.

If any of those happen, you have the right to ask for a human review. That right already exists under Australian administrative law, and the new plan is designed to make it easier to exercise.

Researchers and civil-society groups who have been pressing for these protections for years will be watching closely to see whether the final rules have teeth or stay at the level of guidelines.