Submission 503
Name withheld
Submitter group: Individuals.
Position labels are not shown for individual submitters.
An AI model identified the submitter group and chose the extracts. Counts do not measure importance or merit.
Extracts from the submission, in the submitter’s own words.
Extracts
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.
I think we have reached AGI but ASI is not imminent, and my recommendations come from this.
The committee support a focus on baseline cybersecurity maturity uplift in light of AI threats.
Parliament legislate a requirement to notify an affected individual where AI is materially used in making an administrative decision, and extend the duty to give reasons under the Acts Interpretation Act so it exists regardless of how the decision is made.
The broad and amorphous nature of harms from AI use justifies a generalised obligation for providers to assess and mitigate risks to users, similar to the proposed OSA duty of care.
It has recently shed more than 100 staff and been scrapped as a standalone research unit, alongside deep cuts across CSIRO. I do not think the new AISI can fill this knowledge gap with ongoing funding, per capita, less than a fifth of that of the UK AISI.
In sum, the status quo enables profiting in Australia from misuse of IP, clear solutions based on individual rightsholders are infeasible, and any reform will be first-in-the-world.
Extracts from recommendations
Extracts chosen by an AI model, in the submitter’s own words. Listed in the same order as on its page in the interactive guide.
The committee support a focus on baseline cybersecurity maturity uplift in light of AI threats.
The committee not endorse dedicated investment in AI for cybersecurity.
Parliament legislate a requirement to notify an affected individual where AI is materially used in making an administrative decision, and extend the duty to give reasons under the Acts Interpretation Act so it exists regardless of how the decision is made.
Introduction of law establishing a rebuttable presumption of liability for the legal person responsible for a decision, regardless of whether the decision is automated.
the best remedy is adequate resourcing for the OAIC
Parliament to regulate biometric surveillance in high-risk and legally significant decision making, and apply a moratorium on those uses until commencement.
The EU’s AI Act requirement for machine-readable watermarks seems sensible: R EC OMMENDATION 6 A legislated mandate for watermarking of AI generated content.
A new legislated consumer AI duty of care, similar to the proposed Online Safety Act amendments.
The broad and amorphous nature of harms from AI use justifies a generalised obligation for providers to assess and mitigate risks to users, similar to the proposed OSA duty of care.7 These harms differ from those of other online services, as users relate directly to the service rather than to other people using it. Given eSafety’s existing “systemic workload issues” and failure to apply some OSA provisions, I would place this outside the OSA and eSafety’s remit.
Government responsibility for AI be consolidated in a new independent agency under PM&C.
The Government endorse international efforts to anticipate and avoid catastrophic outcomes from runaway AI. I do not support significant funding or legislative effort on this issue.
An 18-month review of the AI Safety Institute be undertaken in mid-2027.
Consideration of funding and independence comparable to the UK’s AISI, and a clearly distinct scope if the local equivalent is to have an impact.
Adopting a permissive copyright regime with model legislation from the AI companies.
In exchange, imposing a single generative AI levy, on the order of 10% of Australian revenue.
Applying a four year sunset clause to any copyright exemption, preceded by a statutory review.
This should be coupled with ironclad restrictions on operators. The worst demonstrated excesses of data centres, like using illegal fossil fuel generators11 or dumping “bacteria-contaminated water into public sewers” are clearly far below Australian standards.