Parliament’s powerful audit committee will run the ruler over public sector AI use after it was revealed the technology was being deployed widely last year without governance frameworks in place.
The federal government has since issued an official policy for AI use but it has carve outs for Defence and security agencies, and the Joint Committee of Public Accounts and Audit (JCPAA) wants its own oversight.
The JCPAA, which works closely with the national audit office that uncovered the AI use earlier this year, self-referred the inquiry on Thursday.
Unlike other committees, it does not need to set a reporting date but it has asked for submissions by the end of October and committee chair Linda Burney is retiring at the next election.
The Australian National Audit Office last December revealed more than half the federal government agencies it found using artificial intelligence did so without creating policies to govern its use.
When the JCPAA examined the finding in April, then acting auditor general Rona Mellor said AI governance frameworks “weren’t deeply considered or considered [at all]”.
“That’s an area we think people need to work through,” she said at the time.
JCPAA chair Linda Burney said the finding showed more oversight is needed.
“The Committee has decided that oversight is needed of the current and potential future impacts of this fast-developing technology in delivering outcomes for the Australian public,” Ms Burney said.
Terms of reference say the inquiry will examine public sector AI use and the internal and external policies and frameworks in place to support it.
It also has scope for whether the public service has the internal capability to effectively deploy AI and if there are “sovereign capability issues to consider given that most AI tools currently used in Australia are sourced from overseas”.
The Digital Transformation Agency last month issued the first policy for responsible use of AI by government. It came into force this month and offers a principles framework, but agencies are asked to develop their own statements on AI use and risk management.
Several entities testified on their AI use and governance to the JCPAA earlier this year as part of its inquiry into Commonwealth financial statements. This evidence will be used in the new inquiry.
Do you know more? Contact James Riley via Email.