Centre for Public Integrity warning on unregulated government AI
The Centre for Public Integrity has released a report titled “Computer Says No”. It warns that the Albanese government is “sleepwalking into another robodebt scenario” by rolling out algorithms and artificial intelligence across public administration without adequate monitoring. The Guardian reported on the report on 4 November 2026 (5 November 2026 AEST). The thinktank argues that Commonwealth agencies are still deploying automated decision-making (ADM) tools without enough transparency, statutory guardrails or independent auditing, even after the Robodebt Royal Commission exposed the administrative and human cost of automated government decisions.
The trigger for the report is the Department of Health and Aged Care’s Integrated Assessment Tool (IAT). Its algorithmic scoring system was found to systematically under-assess aged care support packages. Human assessors were also explicitly barred from overriding incorrect automated results. Dr Catherine Williams, Chief Executive of the Centre for Public Integrity, said the organisation had become “increasingly concerned about what is clearly the lack of an appropriate and robust framework to govern use of ADM by the government”, and that this was why it decided to put the work forward now.
This is a technology and governance story, not an environmental one. It still matters to Australian environmental professionals and their clients. Developers, lawyers, councils and consultancies increasingly work alongside automated screening tools, risk scoring systems and AI drafting assistants. They all depend on statutory decisions that must be made lawfully by a person with real discretion. The administrative law problem the IAT exposed is not limited to aged care. It applies wherever a decision maker’s judgement is replaced, rather than supported, by a machine output.
Mandatory ADM audits and a public algorithm register
The “Computer Says No” report sets out eight reform recommendations. Two lead the list:
- a statutory requirement for mandatory independent audits of all ADM models, both before and during deployment
- the immediate creation of a publicly searchable register documenting every algorithm used in decisions affecting citizens.
Together these would shift Commonwealth ADM governance away from internal self-assessment and towards external, verifiable scrutiny.
The report also calls for an independent automated decision-making oversight authority with formal audit, inquiry and corrective powers over federal agency tooling. This goes beyond a register. A body with corrective powers could require an agency to change or withdraw a tool, rather than simply report on it. That reflects the view that transparency alone did not prevent Robodebt, and that a watchdog needs enforcement capacity to change agency behaviour.
The evidence base includes a recent Commonwealth Ombudsman review of the IAT. As reported, the review found that agencies bypassed core Robodebt Royal Commission transparency recommendations during procurement and integration of the tool. Appeals against assessments rose more than fivefold. That figure is the clearest measurable sign that the tool’s outputs were not matching the needs of the people being assessed. It also shows how contested automated decisions shift cost and workload into review processes rather than removing it.
On the policy timeline, the Attorney-General’s Department closed consultation on a proposed Commonwealth ADM framework in January 2025. No draft legislation or enforceable timeline has followed. Crossbench MP Kate Chaney has introduced a private member’s bill seeking to impose statutory obligations on government use of ADM. The Robodebt Royal Commission report, delivered in July 2023, recommended a consistent legal framework for automation in government services and a body to monitor and audit automated decision-making. More than three years on, the report argues, neither is in place.
Australian context
The core legal issue raised by the IAT is the fettering of statutory discretion. This is a long-established ground of review in Australian administrative law. Under the Administrative Decisions (Judicial Review) Act 1977 (Cth), an improper exercise of power includes exercising a discretionary power at the direction or behest of another person. It also includes applying a rule or policy without regard to the merits of the particular case. A directive that prevents an assessor from departing from an algorithm’s score raises both concerns. The software effectively becomes the decision maker, and the human signature becomes a formality.
The same principle applies to statutory decisions in contaminated land and environmental regulation. State regulators exercise discretion under legislation such as:
- the Contaminated Land Management Act 1997 (NSW)
- the Environmental Protection Act 1994 (Qld)
- the Environment Protection Act 2017 (Vic)
- the Environment Protection Act 1993 (SA).
Those decisions include issuing notices and orders, determining licence conditions and assessing whether contamination is significant enough to warrant regulation. Statutory site auditors also exercise independent professional judgement when issuing Site Audit Statements or environmental audit statements. Whatever tools support these decisions, the decision must remain one the authorised person actually made, considering the merits of the specific site.
For businesses and professional services more broadly, some transparency obligations for automated decisions are already legislated. Reforms to the Privacy Act 1988 (Cth) passed in late 2024 will require entities to disclose in their privacy policies when personal information is used in substantially automated decisions that significantly affect individuals. Those requirements commence in December 2026. If the government adopts the report’s leading recommendations, mandatory independent audits of ADM models and a public register of algorithms would apply on top of that baseline for Commonwealth agencies.
References and related sources
- Primary source: www.theguardian.com
- indapps.in
- truvace.com
- superpowerdaily.com
How iEnvi can help
iEnvi integrates technology and data-driven approaches into environmental consulting. We monitor AI and technology developments that affect how environmental professionals deliver services to clients.
This is an iEnvi Machete news summary. Prepared by iEnvi to summarise the source article for environmental professionals tracking AI, data, and technology developments that affect consulting and project delivery.
Published: 05 Nov 2026
Need advice on this topic? Speak to an iEnvi expert at info@ienvi.com.au or 1300 043 684, or contact us online.
Need advice on this issue? iEnvi provides practical, senior-led environmental consulting across contaminated land, remediation, ecology and environmental risk.