Statements by Senators: Artificial Intelligence
Parliament of Australia
READ
Details
- Date Published
- 19 Nov 2024
- Priority Score
- 4
- Australian
- Yes
- Created
- 16 Sept 2026, 05:22 am
Description
Search engine for Australian parliamentary documents and media.
Summary
Senator Fatima Payman highlights significant failures in the Australian government's implementation of its own Public Service-wide AI policy, noting that dozens of agencies missed deadlines for transparency statements and accountability appointments. The address exposes instances of agencies providing false or misleading evidence to Senate committees regarding their AI compliance, attributing these failures to administrative errors and lack of oversight. This internal governance struggle raises critical questions about the Australian government's capacity to regulate AI across the broader economy and critical sectors like healthcare and education. While primarily focused on administrative accountability, the findings underscore the difficulty of establishing the robust governance frameworks necessary for managing long-term AI safety and catastrophic risks.
Body
Senator PAYMAN
(Western Australia—Australia's Voice Whip) (13:49): In March, I addressed the chamber on a variety of topics, including the Public Service wide AI policy. Under this policy, 13 agencies missed the deadline to publish an AI transparency statement, and 30 agencies missed the deadline to appoint an AI accountable official. What's more concerning is the number of agencies who gave false evidence to committees of the Senate about this. There are a few, so I'll go through them quickly.
The Office of the Inspector-General of Aged Care gave false evidence to a committee because it 'misunderstood the question'. The Commonwealth Grants Commission misled a committee in relation to both elements of the policy in answers that it admitted were 'provided in error'. The Australian Transport Safety Bureau claimed that their incorrect evidence in relation to both elements was the result of a 'genuine mistake'. The Royal Australian Mint 'mistakenly' advised a committee that it had met the AI transparency statement deadline. The Workplace Gender Equality Agency actually did meet the accountable official deadline but for some reason told a committee that they didn't. One of the agencies wrongly accused of missing the deadline later discovered why. Rather than making a simple Google search, the Digital Transformation Agency responsible for administering this policy had tried to guess their CEO's email address and got it wrong. When the bounce back landed in the DTA's inbox, it went straight to junk mail.
If this government can't even regulate its own use of AI, how can it be expected to regulate AI in our workplaces, our hospitals, our schools right across the country?