Guardian opinion urges federal human rights law to govern AI decisions
The Guardian article argues Australia needs transparency and human review for automated decisions, citing alleged problems with an aged-care assessment tool.
A Guardian opinion article has called for Australia to introduce a federal human rights act requiring transparency and human review of algorithmic decisions.
The article focuses on the automated integrated assessment tool, or IAT, used to determine older Australians’ eligibility and access to aged-care services. It alleges the tool has produced inappropriate outcomes and operates without human oversight or an override option.
According to the opinion article, state health officials compiled spreadsheets of “misaligned” decisions after the tool was introduced. It also cites a warning from WA Health that vulnerable older people could face catastrophic outcomes without a pathway for discretion.
The claims have not been independently established in the supplied material, which includes no figures for affected people, details of the number of decisions involved or response from the Australian Government.
The article argues that existing law does not fully implement recommendations from the robodebt royal commission on disclosure and human review. It also says the Aged Care Act recognises dignity, respect and independence, but that these rights are not fully legally enforceable.
Australia has no federal human rights act, although Victoria, Queensland and the ACT have human rights laws. The article says departments were required to appoint chief AI officers by July and that privacy statements covering AI in consequential decisions were due to be updated by December.
