The Australian Office of the Information Commissioner (OAIC) says that lack of disclosure around use of automated decision-making (ADM) systems across government agencies is putting trust in the technology and the public service at risk.
A desktop study found that only a minority of federal agencies permitted to use ADM technology disclose the practice.
The Office said that more need to do so in order to avoid jeopardising community trust in the government.
The OAIC said that high-profile examples of the technology’s failings, such as those revealed by the Robodebt Royal Commission into Centrelink’s disastrous automated debt recovery scheme, had increased the need for transparency.
“The benefits of utilising ADM technology in government will only be realised if risks are appropriately mitigated and trust is built with the Australian community,” the OAIC said in its report.
“A key enabler of trust is transparency and by leading the way in transparency, the government can build public trust in the technology and ensure its benefits are shared widely across society’,” it added.
Australia’s Information Commissioner Elizabeth Tydd said that the public disclosures were needed for retaining community confidence in government probity.
“By clearly explaining when and how ADM is used, agencies can improve integrity, strengthen public confidence and uphold the statutory right to government information,” Tydd said.
Spanning a range of uses, from simple benefit calculators to machine learning systems, government agencies can use ADMs to determine individuals’ entitlements and obligations.
The systems can benefit the public through increased efficiency and consistency, but also carry serious risks as the devastating political impact of Centrelink’s algorithmic Robodebt debt collection program showed.
Under the transparency requirements of Australia’s Freedom of Information Act, agencies are obliged to tell members of the public of their practices around automated decision-making. However, the OAIC found that few actively do so.
The desktop review found only four agencies disclosed their use of the technology through the Information Publication Scheme (IPS).
The IPS sets standards for public access to government information and requires agencies to publish advice about what they do and the way they do it.
Those agencies were, however, not clear on how they used ADM technology, the OAIC found.
Another nine agencies made reference to use of ADM technology as part of their IPS obligations, but did not confirm if they used it.
The OAIC also found evidence suggesting that another two government entities also used ADM technology.
None of the 23 entities reviewed had published guidelines or policies on how they use the technology, the OAIC further found.
Government use of ADM technology more broadly has been plagued by transparency issues.
Recently, a parliamentary committee raised concerns that ministers might use the technology to, for example, make decision about biosecurity and migration with little justification or explanation made available for public scrutiny.
There is, however, some political will to tackle the lack of transparency around ADM system use.
In response to the Robodebt Royal Commission, the federal Attorney-General’s Department is currently developing a whole-of-government framework, to provide additional guidance on the agencies’ transparency obligations when using ADM and safeguards they have in place to prevent its misuse.
In 2024, Queensland’s whole-of-government ICT oversight and enablement body said that public sector projects would be subject to internal assessments and external reviews to evaluate and mitigate risks related to the use of ADM and artificial intelligence (AI) technologies.