AI Inquiry Atlas

An unofficial, independent guide to the public records of the Joint Select Committee on Artificial Intelligence. Not connected with the Parliament of Australia or the committee. Its makers also made a submission (no. 481; see About). Summaries and scores are written by AI and can be wrong: check the original before relying on them.

Submission 302

Westpac

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.

Open in the interactive guide

Summary

Read the original submission for the full text.

Westpac recommends portable national AI skills and accreditation, nationally consistent principles-based regulation complementing existing law, and cross-border provider accountability. It seeks sovereign AI investment across strategically important parts of the value chain, including compute and infrastructure, while saying this does not mean building every component domestically.

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.

  1. Says workforce capability, rather than access to technology, is now the greatest constraint on AI adoption.

    The greatest constraint on AI adoption is no longer access to technology. It is workforce capability.

  2. Argues for a nationally consistent approach to AI regulation rather than a patchwork of state-based rules inconsistent with international approaches.

    It would be a poor outcome for national productivity if Australia's overarching AI governance model ends up as a patchwork of different state-based AI regulation or rules that are inconsistent and/or unnecessarily diverged from leading international approaches.

  3. Argues that any new AI requirements should be necessary, targeted, proportionate, evidence-based and limited to risks existing frameworks do not adequately address.

    Any new requirements should be necessary, targeted, proportionate, evidence-based and limited to risks not adequately addressed by existing frameworks.

  4. Argues for regulatory traceability of accountability to engine owners where their systems are misused for malintent.

    There must be some regulatory traceability of accountability back to engine owners where their systems are misused for malintent.

  5. Argues that sovereign capability needs a national technology stack across the AI value chain, extending beyond an Australian frontier model.

    This is often referred to as needing an Australian developed ‘AI frontier model’, but it requires much more than that – we need a national technology stack across the AI value chain.

Recommendations

Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.

  1. Consider nationally consistent, portable AI capability through three levels, including standardised AI literacy in school, university and VET curricula.

    Australia should support a nationally consistent approach to AI capability that builds portable skills across organisations and industries and applies the same organising principles from the outset. This could be tiered or considered through three levels of the economy: • Level 1 – Schools and universities. AI literacy delivered through schools, universities and the Vocational Education and Training sector should be developed and standardised in relevant curriculums. Children learning AI fundamentals in a school in Perth, should be consistent with children are leaning in a school in Brisbane.

  2. Consider workforce accreditation providing portable AI skills recognised across employers and sectors, potentially through micro-credentials delivered with private companies, Government and training providers.

    Level 2 – Workforce proficiency. A baseline level of AI accreditation that gives employees the skills and confidence to use AI in their occupations and industries, which would offer employees portable skills recognised across employers and sectors. This could take the form of micro-credentials delivered in partnership between private companies, Government and training providers. There are many good examples that exist today where employers nation-wide recognise an employee’s attained competencies and skills, such as LEAN, MS PowerBI, CompTIA, CPA micro-credentials, WH&S certificates, and many more.

  3. Consider specialist accreditation for professionals designing, deploying, governing, assuring, overseeing and ensuring the safety of AI systems.

    Level 3 – Specialist industry practitioners. Specialist accreditation for professionals responsible for the design, deployment, governance, safety, assurance and oversight of AI systems.

  4. Have Government, industry and education providers build on existing AI micro-credential initiatives to support scalable, nationally consistent capability development.

    Government, industry and education providers should build on these efforts to support scalable, nationally consistent capability development across the economy.

  5. Adopt a nationally consistent approach to AI regulation that complements existing laws and sectoral frameworks and does not duplicate or conflict with existing regulation.

    Adopt a nationally consistent approach to AI regulation that complements existing laws and sectoral frameworks, and which does not duplicate or conflict with existing regulation; and

  6. National Cabinet should maintain a national cross-regulatory AI implementation roadmap mapping AI obligations across states, territories and regulators and identifying overlaps, inconsistencies or gaps.

    National Cabinet should maintain a national cross-regulatory AI implementation roadmap that maps the principal regulatory obligations relevant to AI across states, territories and regulators, and identifies areas of overlap, inconsistency or potential regulatory gaps.

  7. Align Australia's national AI framework with international approaches to support businesses operating in global markets.

    Australia’s national framework should be aligned with international approaches to support Australian businesses operating in global markets.

  8. Maintain principles-based, technology-neutral AI regulation proportionate to risk, continuing to avoid needlessly prescriptive or technology-specific rules.

    Government should maintain its commitment to principles-based and technology-neutral AI regulation that is proportionate to risk, with obligations calibrated to the potential for harm. Government should also continue to avoid needlessly prescriptive or technology specific rules that are likely to quickly become outdated or unnecessarily constrain beneficial uses of AI.

  9. Make new AI requirements necessary, targeted, proportionate, evidence-based and limited to risks existing frameworks inadequately address; avoid an AI-specific overlay where existing obligations suffice.

    Any new requirements should be necessary, targeted, proportionate, evidence-based and limited to risks not adequately addressed by existing frameworks. Government should continue its current approach of avoiding an additional AI specific regulatory overlay where existing obligations are already sufficient.

  10. Invest in regulators’ capability for emerging AI risks and opportunities; give the Australian AI Safety Institute a complementary role focused on shared understanding of emerging risks, standards and good practice.

    Government should also invest in the capability of regulators to keep pace with emerging AI risks and opportunities. The Australian AI Safety Institute should complement, rather than duplicate, existing regulators, with a clear role focused on shared understanding of emerging risks, standards, and good practice.

  11. Consider how liability and enforcement arrangements will operate across global jurisdictions.

    Policymakers should also consider how liability and enforcement arrangements will operate across global jurisdictions.

  12. Seek in future AI governance to ensure technology providers take appropriate responsibility for the impacts of their products and services in Australia, regardless of where they are based.

    Much like Australia’s efforts to regulate social media platforms and establish clear expectations for their operation in the Australian market, future AI governance arrangements should seek to ensure that technology providers take appropriate responsibility for the impacts of their products and services within Australia, regardless of where those providers are based.

  13. Require greater transparency from frontier model developers on model capabilities, limitations, known vulnerabilities and material changes between versions.

    Effective management of these risks will require greater transparency from frontier model developers regarding model capabilities, limitations, known vulnerabilities and material changes between versions.

  14. Continue to examine how accountability is shared across the AI value chain, including the responsibilities of model developers and platform providers where foreseeable misuse arises from capabilities within their control.

    Policymakers should continue to examine how accountability is shared across the AI value chain, including the responsibilities of model developers and platform providers where foreseeable misuse arises from capabilities within their control.

  15. Provide appropriate mechanisms to ensure responsibility for downstream consequences is not borne solely by those deploying AI.

    As organisations are increasingly expected to manage the downstream consequences of these technologies, there should be appropriate mechanisms to ensure responsibility is not borne solely by those deploying them.

  16. Provide regulatory traceability of accountability back to owners of AI engines where their systems are misused for malintent.

    There must be some regulatory traceability of accountability back to engine owners where their systems are misused for malintent.

  17. Ensure policy settings let Australian AI businesses start, scale and compete while meeting procurement, security and outsourcing obligations of regulated sectors.

    It is in the national interest to ensure that policy settings create an environment where Australian AI businesses can start, scale and compete successfully while continuing to meet the procurement, security and outsourcing obligations required of regulated sectors.

  18. Build sovereign AI capability through a national technology stack, including cloud and data centres, chips and technology capabilities and domestic compute as foundations for domestic frontier models.

    While growing a strong base of domestic AI services is important, the real opportunity for Australia’s long-term economic and technology resilience will come from developing a home-grown technology ecosystem which supports local AI development. This is often referred to as needing an Australian developed ‘AI frontier model’, but it requires much more than that – we need a national technology stack across the AI value chain. This includes building out Australia’s cloud and data centre infrastructure, chips and technology capabilities, harnessing our own compute power, and only then will we have the foundations to grow the development of domestic frontier AI models. We acknowledge this is easier said, than done.

  19. Consider sovereign AI capability alongside national priorities such as defence, energy, critical minerals and housing.

    Australia already invests in nationally significant capabilities and infrastructure in areas such as defence, energy, critical minerals and housing. As AI becomes increasingly important to productivity, innovation and economic resilience, sovereign AI capability should be considered alongside these national priorities.

  20. Consider how the Future Fund, Business Growth Fund and National Reconstruction Fund can align investor return objectives with sovereign AI goals.

    Australia should consider how existing investment settings, including the Future Fund, Business Growth Fund and National Reconstruction Fund, can align existing investor return objectives with broader national sovereign AI goals.

  21. Place public-private collaboration and investment in R&D, sovereign compute, nationally significant infrastructure and AI-value-chain skills at the centre of Australia’s long-term AI strategy.

    We will require an ambitious program of public and private collaboration and investment to support research and development, sovereign compute capability, nationally significant infrastructure, and the skills needed to participate across the AI value chain. Building these capabilities should sit at the centre of Australia’s long-term AI strategy.

Read the original

Open in the interactive guide