The Centre for Public Integrity released a report on 5 November calling for a mandatory statutory framework governing automated decision-making across the Australian government.
Computer Says No finds that agencies increasingly use automated systems to administer legislation and determine services, but disclosure of where those systems operate—and whether their outcomes are lawful and fair—remains inadequate. The Centre recommends an audit of existing systems, a public register, independent oversight, parliamentary authorisation for categories of automated statutory decisions, pre-deployment legal scrutiny, continuing bias testing and enforceable rights to reasons and review.
These are recommendations, not law or government policy. A private member’s Automated Decision-Making (Safeguards and Transparency) Bill 2026, introduced on 7 September, remains before the House of Representatives.
The report matters because it treats automation as an administrative-law problem, not merely an AI-safety issue. Its proposed controls focus on legal authority, responsibility, evidence and contestability throughout a system’s lifecycle.
For lawyers advising government or technology suppliers, the practical implications go beyond checking model accuracy. Procurement terms, decision logs, human-override arrangements and explanations supplied to affected people may determine whether an automated process is lawful, reviewable and defensible.
Source: Centre for Public Integrity
Leave a Reply