The Algorithmic Shadow: Why Australia’s Automated Decisions Demand Immediate Transparency

• The Centre for Public Integrity warns the government is "sleepwalking" into a crisis similar to the robodebt scandal due to insufficient transparency in algorithmic decision-making.
• A new report calls for mandatory audits, a public register of all ADM usage, and an independent oversight body with investigative powers.
• Concerns are specifically heightened regarding the Integrated Assessment Tool (IAT), which routinely under-assesses aged care needs while banning assessors from overriding errors.
• The Attorney General’s department has been developing an ADM framework since before January 2025, though no draft has yet been made public.
• Independent MP Kate Chaney introduced a private member’s bill to establish a legislated, mandatory framework for government ADM use.
• Critics argue that current automation efforts prioritize cost savings over the transparency and trust required for fair administrative processes.

The core of the controversy lies in the lack of public knowledge regarding whether algorithms influence decisions that affect individual lives. According to Dr. Williams, this opacity is particularly dangerous given the history of automated systems failing vulnerable populations. The Centre for Public Integrity has identified eight key recommendations in their report, all aimed at closing these governance gaps. These include the establishment of a mandatory audit process and a public register that tracks every instance of ADM use by government entities. Furthermore, they advocate for an independent oversight body equipped with the authority to monitor, audit, and investigate government actions related to automated decisions.

A primary catalyst for this urgent call to action is the Integrated Assessment Tool (IAT). This algorithmic system is currently used to determine the level of home support that older Australians are entitled to receive. Investigations have revealed that the IAT routinely under-assesses individuals' needs, resulting in lower package entitlements than they should receive. Compounding the issue, government regulations prohibit aged care assessors from overriding these incorrect algorithmic decisions, leaving citizens with no recourse within the standard assessment process.

"We have been watching the evidence emerge in relation to the aged care algorithm tool, and becoming increasingly concerned about what is clearly the lack of an appropriate and robust framework to govern use of ADM by the government," Williams stated. She emphasized that the situation feels like a repetition of past mistakes, noting, "It feels as though we are sleepwalking into another robodebt scenario." This sentiment drove the decision to publish their findings now, with the hope that it might inform any future frameworks the government proposes regarding ADM governance.

In response to these growing concerns, the Attorney General’s department has been working on an ADM framework. Public consultation for this initiative concluded in January 2025. However, Attorney General Michelle Rowland did not provide answers regarding when a draft of this framework would be completed or made public. Instead, she issued a statement affirming that the government "is committed to ensuring the safe and responsible use of ADM to deliver timely and efficient services for the Australian community."

Rowland noted that the upcoming framework would implement recommendations from the royal commission into robodebt. She claimed the new guidelines will be designed to improve the accuracy and consistency of administrative decisions, embed robust safeguards including transparency obligations, and ensure that ADM is used lawfully with appropriate human oversight. Additionally, she argued that these measures will lift productivity and help deliver faster, more reliable services.

Despite these assurances, a report from the Commonwealth Ombudsman published in October revealed that transparency recommendations from the robodebt royal commission had been ignored by the government during the development and rollout of the IAT. This finding underscores the disconnect between official promises and operational reality.

The lack of progress in formalizing ADM rules has prompted action from independent lawmakers. In September, Independent MP Kate Chaney introduced a private member’s bill aimed at establishing the first legislated, mandatory framework for government use of ADM. Chaney highlighted that the Attorney General’s department had finished its consultation process on an ADM framework almost two years ago.

"I have met with the attorney general about it and they say ‘We’re working on it’," Chaney recounted. She stressed the importance of maintaining pressure on the government to ensure the initiative is not abandoned due to complexity. "I think it’s really important that we keep the pressure up so that it doesn’t get dropped because it’s too hard."

Chaney pointed out that various government departments are currently making decisions about ADM use with little consistency or transparency. She suggested that while automating decision-making can be a method for achieving cost savings in the public service, it must not come at the expense of public trust. "There does need to be transparency about when decisions are automated, and when it’s not appropriate for those to be automated," she said. Citizens require clear understanding of why decisions are made and clear pathways to review them. "This is what people expect of government."

The debate over automated decision-making in Australia has moved beyond technical implementation to fundamental questions of accountability and trust. With the Centre for Public Integrity warning of a potential repeat of the robodebt disaster, and independent MPs pushing for legislative mandates, the pressure is mounting on the federal government to provide more than just assurances. The demand is for concrete, transparent, and legally binding frameworks that ensure algorithms serve the public interest rather than obscure administrative errors. Until such safeguards are in place, the risk of automated systems undermining fair treatment remains high.