Australian Ombudsman Urges Human Override for Aged-Care Funding Algorithm
An Australian oversight investigation says automated aged-care assessments lack adequate safeguards and transparency, prompting calls to let clinicians correct inappropriate funding decisions.
Australia's Commonwealth Ombudsman has called for changes to an automated system that determines the level of government-funded home support available to older people, finding shortcomings in transparency, oversight and the ability to correct unsuitable decisions.
The findings, released on October 26, concern the Integrated Assessment Tool (IAT), which forms part of the national Single Assessment System. The tool helps classify applicants for home-based aged-care services and associated funding. It was introduced in November 2025.
Ombudsman Iain Anderson found that the legal framework does not allow assessors to override an algorithmic outcome even when it is clinically inappropriate. The investigation also criticised limited public information about how the tool converts assessment responses into decisions, along with weaknesses in supervision and verification of assessor qualifications.
The report made 10 recommendations, including legal changes to allow appropriately qualified staff to correct erroneous automated outcomes and a wider review of the assessment framework. It also called for meaningful public consultation on future changes.
The government accepted six recommendations and noted others, including the proposed human-override mechanism, according to ABC News. Ministers Mark Butler and Sam Rae said the government would examine the findings and continue work to make the aged-care assessment system more transparent and reliable. The department disputed the Ombudsman's characterisation of its testing and evaluation processes.
An escalation pathway for exceptional cases has been introduced, but it is distinct from granting frontline assessors general authority to change the algorithm's funding classifications. A further independent review is expected in the first half of 2027.
The dispute concerns the governance of automated public-service decisions rather than a finding that every assessment is incorrect. The central question is how to ensure that funding classifications can be explained, challenged and corrected when individual care needs are not accurately reflected.
AI assistance was used in preparing this report. Fast News remains responsible for its editorial verification and publication.
Spotted an error? Request a correction · Our editorial principles · Contact the newsroom