Aged Care Legislation Amendment (Aboriginal and Torres Strait Islander Aged Care Commissioner and Other Measures) Bill 2026

Ms CHANEY (Curtin) (10:19): Among other things, the government's amendments to the Aged Care Act introduce an additional human review process for decisions made by the integrated assessment tool, and I support these amendments. They're overdue and they're welcome. For close to a year, an automated tool has had the final say on aged-care support packages for older Australians.

Assessors with decades of clinical experience have been unable to override a result that they believed to be wrong. Older Australians have seen their supports cut despite deteriorating health, with no clear explanation and no way for a human to step in and fix it. I've heard directly from individuals affected and from aged-care assessors and providers in Curtin who describe watching this system produce outcomes they knew to be incorrect and being powerless to change them.

That's older Australians carrying the cost of a process that nobody can properly explain or correct. So, when the government brings forward a schedule that restores a form of human override, I welcome it. The new part 4A allows a human assessor to request an additional review of a decision made by the automated integrated assessment tool.

The substance of this reform sits in the rules that will be made under it. The bill leaves crucial matters to those rules, including what requirements must be satisfied for this human review process to be used. Those rules are not before the House today.

They will be made by the executive, and they deserve the same scrutiny as the legislation itself. I'll be looking closely at them when they're tabled, and I expect the government to consult properly with the sector first. A human override with narrow rules and short windows would not fix this.

I also want to place these amendments in a wider context, because it should not have taken this long, and it should not happen again the way it's happened this time. This morning, the Guardian revealed that, before the rollout of the automated aged-care tool, assessors had been told by government many times that they would maintain the ability to override the assessment tool.

Only at the last moment did these assessors, many of whom had been working to assess and support older Australians for decades, realise that the government had actually failed them and failed older Australians by taking away any ability to override the automated tool. Despite frantic warnings made by the assessors that this would cause real problems, the government ignored them and ploughed ahead.

These types of safeguards should not be an afterthought, and they should not be at the whim of each minister. Last week, I introduced the Automated Decision-Making (Safeguards and Transparency) Bill into this House. That bill does for all automated government decisions what these amendments do for one small corner of the aged-care system.

It requires any department or agency automating a decision to meet a set of safeguards and transparency requirements before that automation goes live. It prohibits automation where a decision requires genuine human discretion, where an outcome cannot be properly explained or where it would produce an unlawful result. Where automation is used for high-risk decisions, it requires a human to be able to override the outcome, requires the system to be tested for accuracy before use and requires review pathways and independent oversight.

The story of the integrated assessment tool shows why this framework is needed. This was not unique to aged care. It was a symptom of a larger gap.

We have no general rule that a human must be able to correct an automated government decision. So, when something goes wrong, the fix has to be legislated, department by department, after the harm has already been done. Older Australians should not have to be the ones who discover the gap in the system by living through it.

Neither should people relying on the NDIS or social security or on the growing number of government processes where automation is being introduced. That's the case for a single mandatory framework that applies across government, rather than a patchwork of amendments introduced after each failure has already occurred. I support these amendments.

I'll be scrutinising the rules that follow them closely, and I urge the government to look at what happened here and support a broader framework governing the use of automated decision-making so the next robodebt, no matter which department it's in, is stopped before it starts.