Submission 490
Ms Robyn Sirr
Submitter group: Individuals.
Position labels are not shown for individual submitters.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
A retired health professional argues that Parliament inadequately addresses frontier AI’s catastrophic and loss-of-control risks. Recommends scaling up AISI, focusing on powerful models’ development, supporting an international pause on recursive self-improvement until safety is assured, and parliamentary advocacy. Acknowledges existing useful AI applications and potential wellbeing benefits while urging action on severe risks.
Key points
Quotes are the submitter’s own words. Text taken from documents, including scanned pages read by text-recognition software, can differ from the PDF. Check the original. Listed in the same order as on the submission’s page in the interactive guide.
Argues that risks of harm from AI at a global and catastrophic scale are not being adequately addressed in Parliament.
I’m very concerned that risks of harm posed by AI on a global and catastrophic scale – not just to Australians, but to all humanity – are not being adequately addressed in our parliament.
Argues that AI development is outpacing safety, citing the International AI Safety Report 2026.
AI development is outpacing safety. This has been increasingly cited by experts, such as the International AI Safety Report 2026, contributed to by over 100 scientists from 30 countries including Australia.
Argues that competition between US and Chinese frontier AI companies is driving the pace of development towards superintelligence.
The escalating pace of development is driven by competition between US and Chinese frontier AI companies. The explicit goal of these companies is to develop artificial superintelligence (ASI) - which refers to AI that can substantially outperform humans at all cognitive tasks.
Argues that Australia's AI Safety Institute must be resourced for the magnitude of AI risk, stating the UK funds its institute six to seven times as much per capita.
Australia’s AISI must be resourced for the magnitude of risk that AI poses. On a per-capita basis the UK funds its AISI six to seven times the level of ours, enabling retention of the essential technical expertise and the compute power needed for safety testing of AI models.
Argues that the AISI should focus on risks from the most powerful models being developed in the US and China, not only deployment risks in Australia.
Australia’s AISI should have a strong focus on the risks posed by the development of the most powerful models in the US and China. It is not enough to limit its focus merely on the present-day risks posed by the deployment of AI in Australia.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Listen to the interview on managing AI loss-of-control risk.
He added that an AI-instigated catastrophe could occur in too many ways to speculate, and that regardless, we have to figure out how to manage this loss of control risk before it’s too late. This interview is succinct, highly relevant to the Enquiry: I urge you to give it a listen here:
Scale up and appropriately resource Australia's AI Safety Institute.
Australia’s AISI must be resourced for the magnitude of risk that AI poses. On a per-capita basis the UK funds its AISI six to seven times the level of ours, enabling retention of the essential technical expertise and the compute power needed for safety testing of AI models.
Have the AI Safety Institute focus strongly on risks from development of the most powerful models in the US and China, not only on deployment risks in Australia.
Australia’s AISI should have a strong focus on the risks posed by the development of the most powerful models in the US and China. It is not enough to limit its focus merely on the present-day risks posed by the deployment of AI in Australia.
Call for an international pause on AI recursive self-improvement until safety can be assured.
Australia should add its voice to the call for an international pause on the development of AI recursive self-improvement until we can be assured of its safety.
Speak out as parliamentarians on frontier AI risks.
3. Speak out as parliamentarians Even when their colleagues are slow, parliamentarians in other jurisdictions are responding. Recent voices warning of the risks posed by superintelligence and calling for government intervention include:
Examine loss-of-control risk and advocate a Government response commensurate with the security risk.
I urge the Committee to examine this aspect of loss of control risk, and to advocate for a Government response that is commensurate with the security risk it presents.