Submission 84
Vault
Submitter group: Companies.
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.
Vault, a provider of sovereign AI cloud infrastructure, argues Australia should treat AI infrastructure as a strategic capability, with sovereign, secure infrastructure for sensitive workloads and procurement that strengthens demand for Australian capability. States it does not propose a broad new regime, but AI may expose gaps needing targeted reform; argues tax settings disadvantage domestic operators.
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 Australia should treat AI infrastructure as a core part of its economic and digital foundation, with procurement prioritising security, sovereignty, interoperability and resilience.
Australia should treat AI infrastructure as a core component of its economic and digital foundation. While relying on global technology markets, Australia requires domestic infrastructure to securely operate sensitive AI workloads. Government procurement must therefore prioritise security, sovereignty, interoperability, and resilience alongside cost and performance.
Argues that sovereign AI capability should include the ability to securely develop, deploy and operate AI systems in Australia, not only domestic models.
Australia's sovereign AI capability should extend beyond developing Australian AI models. A sovereign AI capability also requires the ability to securely develop, deploy and operate AI systems in Australia.
Says data sovereignty for AI involves where data is processed, who can access it and who operates the infrastructure, not only physical location.
For AI, data sovereignty is not simply a question of where a database is physically located. Organisations also need to understand where data is processed, who can access it, where metadata and logs are stored, who operates the infrastructure and what legal obligations may apply to the provider.
Argues that regulation should be proportionate, use-case based and technology-neutral, with clear accountability for higher-risk applications.
Regulation should remain proportionate, use-case based and technology-neutral, with clear accountability for higher-risk applications.
Argues that transparency for agencies should focus on assurance about infrastructure rather than disclosure of proprietary technology.
Transparency, on the other hand, should focus on assurance rather than disclosure of proprietary technology. Accordingly, agencies should be provided with appropriate assurance that the infrastructure supporting an AI system meets applicable security, sovereignty and operational requirements
Argues that, absent decisive government intervention, it is highly probable that within a decade Australia's national security and economic stability will fundamentally rely on foreign-controlled AI.
Absent decisive government intervention, it is highly probable that within a decade, Australia's national security and economic stability will become fundamentally reliant on foreign-controlled artificial intelligence.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Make government procurement prioritise security, sovereignty, interoperability and resilience alongside cost and performance.
Government procurement must therefore prioritise security, sovereignty, interoperability, and resilience alongside cost and performance.
Use procurement and investment settings to strengthen demand for, and development of, Australian AI infrastructure including compute and cloud capacity.
The Government can support sovereign AI capability through procurement settings that strengthen demand for Australian technology capability, and through investment settings that encourage the development of Australian AI infrastructure, including compute and cloud capacity.
Ensure access to appropriately sovereign infrastructure for sensitive data and AI workloads, with stronger sovereignty requirements where sensitivity, risk or strategic importance warrants.
The policy objective should be to ensure that Australian organisations have access to appropriately sovereign infrastructure for sensitive data and AI workloads, with stronger sovereignty requirements applying where the sensitivity, risk or strategic importance of the workload warrants it.
Give agencies appropriate assurance that the infrastructure supporting an AI system meets applicable security, sovereignty and operational requirements, with transparency focused on assurance rather than disclosure of proprietary technology.
Transparency, on the other hand, should focus on assurance rather than disclosure of proprietary technology. Accordingly, agencies should be provided with appropriate assurance that the infrastructure supporting an AI system meets applicable security, sovereignty and operational requirements
For higher-risk government AI uses, apply proportionate controls warranted by sensitivity and consequences, covering data handling, infrastructure security, access controls, monitoring, auditability and operational resilience.
For higher-risk applications, existing government requirements should be applied proportionately, with stronger controls where warranted by the sensitivity and consequences of the use case. These should address data handling, infrastructure security, access controls, monitoring, auditability and operational resilience.
Ensure agencies have access to infrastructure that lets them meet their security, privacy and accountability obligations without unnecessarily restricting adoption of leading AI technologies.
agencies should have access to infrastructure that enables them to meet their security, privacy and accountability obligations without unnecessarily restricting their ability to adopt leading AI technologies.
Keep AI regulation proportionate, use-case based and technology-neutral, with clear accountability for higher-risk applications.
Regulation should remain proportionate, use-case based and technology-neutral, with clear accountability for higher-risk applications.
For higher-risk AI deployments, have regulation and assurance frameworks give greater clarity on responsibility for underlying infrastructure, including data location, administrative access, operational control, supply-chain dependencies and incident response.
For higher-risk AI deployments, regulation and assurance frameworks should provide greater clarity around responsibility for the underlying infrastructure. This should extend beyond the AI system itself to include data location, administrative access, operational control, supply-chain dependencies and incident response, consistent with emerging international approaches to cloud and AI sovereignty.
Where organisations already operate within established security and assurance frameworks, use them wherever possible for AI infrastructure confidence, avoiding unnecessary regulatory complexity.
There is also an opportunity to make greater use of existing Australian assurance mechanisms rather than creating parallel regimes. Where organisations already operate within established security and assurance frameworks, these should be leveraged wherever possible to provide confidence in the infrastructure supporting AI. The objective should be clear accountability and proportionate assurance, without creating unnecessary regulatory complexity that could inhibit responsible AI adoption.
Apply stronger security, resilience and operational-control requirements to AI workloads supporting critical infrastructure, sensitive government functions or other strategically important capabilities.
AI workloads supporting critical infrastructure, sensitive government functions or other strategically important capabilities should be subject to stronger requirements for security, resilience and operational control.
Equip regulators and the national security community to identify AI dependencies and disruption impacts, informing risk monitoring and responses to compromised access, availability or control.
For Australia, this means regulators and the national security community need the capability to identify material dependencies across AI compute, cloud, software, data and supply chains, and understand how their disruption could affect critical capabilities. This should inform how emerging AI risks are monitored and how Australia prepares to respond when access, availability or control is compromised.
Act now to build domestic infrastructure, assurance and technology options needed to adopt AI with confidence while remaining connected to global technology markets.
Australia should act now to build the domestic infrastructure, assurance and technology options needed to adopt AI with confidence while remaining connected to global technology markets.