Submission 402
Global Data Alliance (GDA)
Submitter group: Industry bodies.
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.
Argues that sovereign AI capability and data sovereignty should rest on capability, collaboration, resilience and trust, assessed through outcomes rather than nationality or data location. Recommends investment in skills and infrastructure, global collaboration and open data flows. Allows that risk assessments may justify location or access restrictions for classified or highly sensitive government information. The coalition says its members rely on cross-border data access.
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.
The GDA argues that sovereign capability should be distinguished from technological isolationism or exclusion.
In that regard, we encourage the Committee to distinguish between sovereign capability — the practical ability to govern, secure, operate and recover AI-enabled systems, while developing domestic skills, research and commercial capacity — and technological isolationism or technological exclusion.
A program corrected obvious text-recognition errors in this quote for display.
The GDA argues that for data sovereignty, effective control does not generally require exclusive domestic storage or processing.
For data sovereignty in particular, effective control does not generally require exclusive domestic storage or processing.
The GDA argues that Australian firms need access to globally distributed inputs, and that restricting them by nationality or ownership can raise entry costs and hinder commercialisation.
To pursue that opportunity, Australian researchers and firms need reliable access to complementary inputs that are often globally distributed: accelerated compute, base and foundation models, cybersecurity services, specialised enterprise software, and research partners. Restricting these inputs by nationality or ownership can raise entry costs and make it harder for Australian firms to achieve commercialisation.
Describes regional modelling that excluded Australia as finding non-linear increases in costs and risks under origin-based restrictions, supporting caution.
Modelling five levels of AI sovereignty restrictiveness across the Asia-Pacific, the study found that costs and risks rise non-linearly when origin-based restrictions are imposed in place of functional performance-based criteria. While Australia was not included in the study, the regional results nevertheless support caution.
A program corrected obvious text-recognition errors in this quote for display.
The GDA argues that tested continuity and portability, rather than provider nationality, are better indicators of sovereign resilience.
Accordingly, the GDA encourages the Committee to treat tested continuity, redundancy, recoverability, portability, and supplier-risk management as more meaningful indicators of sovereign resilience than nationality or origin.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Define data sovereignty through effective control—access, security, governance, auditability and accountability—rather than exclusive Australian storage or processing.
For purposes of Term of Reference (f), data sovereignty should principally mean effective control over data —including access, security, governance, auditability and accountability — rather than a requirement that data be stored or processed exclusively within Australia.
Strengthen AI infrastructure, AI innovation and AI skills (sovereignty through capability).
Sovereignty through Capability: Strengthen AI infrastructure, AI innovation, and AI skills;
A program corrected obvious text-recognition errors in this quote for display.
Promote sovereignty through industry-to-industry, industry-to-government and government-to-government collaboration across AI domains.
Sovereignty through Collaboration: Promote sovereignty via industry-to-industry, industry-to-government, and government-to-government collaboration across AI domains;
A program corrected obvious text-recognition errors in this quote for display.
Build resilience and redundancy into Australia's AI ecosystem.
Sovereignty through Resilience: Build resilience and redundancy into Australia’s AI ecosystem; and
A program corrected obvious text-recognition errors in this quote for display.
Promote trust in AI governance while preserving the open, cross-border data flows that support AI development, deployment and integration.
Sovereignty through Trust: Promote trust in AI governance while preserving the open, cross-border data flows that support AI development, deployment, and integration across the economy.
A program corrected obvious text-recognition errors in this quote for display.
Invest in human capital, technical infrastructure, R&D and international technology partnerships to strengthen capability.
Recommendation I: Sovereignty Through Capability Capability is a foundation of long-term digital sovereignty. Governments can strengthen it by investing in human capital, technical infrastructure, R&D, and international technology partnerships.
Prioritise AI and cybersecurity skills, high-performance compute, cloud and data engineering, R&D, standards, model evaluation, data curation and commercialisation pathways for Australian research and start-ups.
For AI, priority areas include AI and cybersecurity skills; access to high-performance compute; cloud and data engineering; research and development; standards expertise; model evaluation; data curation; and pathways that help Australian research and start-ups reach commercial scale.
A program corrected obvious text-recognition errors in this quote for display.
Make Australia an attractive place to conduct research, deploy engineers, establish capability centres, develop products and scale firms.
Policy should build on that momentum by making Australia an attractive place to conduct research, deploy engineers, establish capability centres, develop products, and scale firms.
Ensure any recommendations on sovereign capability do not compromise cross-border access to cloud capacity, knowledge and data, sectoral AI integration, upskilling and engagement on interoperable governance.
Broadly speaking, development of such capabilities depends on factors including: * Cross-border access to globally distributed cloud computing capacity; * Cross-border access to knowledge, information, and data; + Integration of AI and other transformative tools into finance, health, manufacturing, and other sectors; * Digital upskilling for researchers, programmers, engineers, technicians, and other professionals; and * Engagement with governments and standards organizations on interoperable governance frameworks. We urge the Committee to ensure that any recommendations relating to sovereign capability do not compromise these success factors.
A program corrected obvious text-recognition errors in this quote for display.
Explore approaches that support interoperability, portability and competitive access, so Australian firms can choose among and change providers.
Government-to-industry collaboration can explore approaches to support interoperability, portability and competitive access. Australian firms should be able to choose among technologies and providers, combine domestic and international inputs, and change providers when business, security or resilience needs require it.
Pursue commitments against arbitrary, discriminatory, disguised or unnecessary restrictions on digital trade, cross-border data transfers or continued technology access.
As an illustration of government-to-government collaboration, governments benefit from commitments to avoid arbitrary, discriminatory, disguised, or unnecessary restrictions on digital trade, cross-border data transfers, or continuity of access to technology.
Keep Australia's policy settings predictable, risk-based and non-discriminatory, rather than broad requirements that AI infrastructure, models or providers be domestically owned or headquartered.
Australia's policy settings should therefore remain predictable, risk-based and non-discriminatory. Broad requirements that AI infrastructure, models or service providers be domestically owned or headquartered can work against these collaborative objectives.
A program corrected obvious text-recognition errors in this quote for display.
Use outcome-based requirements such as security, auditability, continuity, portability, access control and legal accountability to test whether a service advances Australian interests.
Outcome-based requirements - security, auditability, continuity, portability, access control and legal accountability — are better suited to testing whether a service advances Australian interests.
Assess Australia-specific trade-offs before adopting broad origin-based sovereignty requirements.
These results illustrate the types of trade-offs — including higher costs, slower AI diffusion and lower productivity — that warrant careful Australia-specific assessment before broad origin-based sovereignty requirements are adopted.
A program corrected obvious text-recognition errors in this quote for display.
Keep resilience requirements proportionate to workload risk and importance, recognising that organisations need different architectures.
Not every organisation will need every architecture, and resilience requirements should remain proportionate to the risk and importance of the workload.
Treat tested continuity, redundancy, recoverability, portability and supplier-risk management as indicators of sovereign resilience.
Accordingly, the GDA encourages the Committee to treat tested continuity, redundancy, recoverability, portability, and supplier-risk management as more meaningful indicators of sovereign resilience than nationality or origin.
Strengthen sovereign capability by improving the quality, governance, provenance and availability of Australian datasets in areas such as agriculture, resources, medical research, advanced manufacturing and financial services.
Australia can strengthen sovereign capability by improving the quality, governance, provenance and availability of Australian datasets in areas such as agriculture, resources, medical research, advanced manufacturing and financial services.
Require appropriate permissions, documentation, access controls, provenance, security and auditability for trustworthy use of Australian datasets.
Trustworthy use of those datasets should include appropriate permissions, documentation, access controls, provenance, security and auditability.
Outside risk-justified restrictions for classified or highly sensitive government information, assess data control, verifiable security, enforceable accountability and operational control.
For classified military, intelligence, or other highly sensitive government information, a specific risk assessment may justify additional location or access restrictions. Outside those narrow circumstances, however, the better test is whether Australia can control access, verify security, enforce accountability and maintain operational control.
Read the original
In some quotes a program corrected obvious text-recognition errors for display.