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 431

Mr Raffaele D'Angelo

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.

Proposes an Australian AI National and Community Benefit Framework linking adoption, sovereignty, infrastructure, safety and skills to measurable public value, with recommendations on partnerships, assurance, Commonwealth governance, provenance, dependencies and community benefits. The author discloses advisory work for Basketball NSW and overlapping interests, and states these are personal views, not that organisation’s.

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. Argues that AI success should be measured by enduring Australian capability, productivity, resilience and community value rather than technology hosted or consumed.

    Australia should measure AI success not by how much technology it hosts or consumes, but by how much enduring Australian capability, productivity, resilience, and community value is created.

  2. Argues that SMEs, not-for-profits and community organisations often cannot afford compute, legal and governance capability, and that scale-dependent AI may widen capability gaps.

    Large organisations can purchase compute, legal advice, cyber capabilities and specialist governance. SMEs, family businesses, not-for-profits and community organisations often cannot. If safe and effective AI depends on scale, it may widen existing capability gaps.

    A program corrected obvious text-recognition errors in this quote for display.

  3. Argues that sovereignty depends on more than hosting location, including ownership and control, how data is routed and stored, and which jurisdictions and subprocessors apply.

    Physical hosting in Australia can strengthen resilience, but it does not establish sovereignty on its own. Relevant questions include ownership and control of infrastructure and models; where data is routed, cached, backed up and analysed; which jurisdictions and subprocessors apply

  4. Argues that Commonwealth AI should set a governance benchmark: agencies identifying purpose, accountable owner, data sources, risk classification, oversight and review for material systems.

    The Commonwealth can set a practical benchmark. For material AI systems, agencies should be able to identify purpose, accountable owner, lawful basis, data sources, risk classification, evaluation, human oversight, incidents, significant vendors, dependencies and review arrangements.

  5. Says Australia may become highly dependent on a small number of global providers, and that government should understand substitutability and exit arrangements.

    Australia may become highly dependent on a small number of global cloud, model and infrastructure providers. Government should understand substitutability, portability, exit arrangements, supply-chain dependencies, foreign legal exposure and recovery capability before dependencies become difficult to unwind or recalibrate.

  6. Argues that major AI and data centre infrastructure should be assessed by where investment can create the greatest enduring public value, not only by whether a site can accommodate it.

    Major AI/data-centre infrastructure should be assessed not only by whether a site can accommodate it, but by where the investment can create the greatest enduring public value.

Recommendations

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

  1. Establish an Australian AI National & Community Benefit Framework linking productivity, adoption, sovereign capability, data sovereignty, infrastructure, safety and skills to governance and measurable public value.

    Australia should establish an Australian AI National & Community Benefit Framework that links productivity, adoption, sovereign capability, data sovereignty, infrastructure, safety, and skills to governance and measurable public value.

  2. Create sector-based public-interest AI adoption partnerships, including a sport/community pilot combining compute access, governance, training, data capability and measurable outcomes.

    Create sector-based public-interest AI adoption partnerships including a sport/community pilot that combines computer access, governance, training, data capability and measurable outcomes.

    A program corrected obvious text-recognition errors in this quote for display.

  3. Support adoption through proportionate governance tools, training for boards, executives, workers and volunteers, procurement/evaluation guidance, expertise and compute access, sector use cases and controlled tests, and measurement of benefits, failures and unintended consequences.

    2.2 A public-interest adoption model = simple, proportionate governance and assurance tools; = training for Boards, executives, workers and volunteers; = shared guidance on procurement and model evaluation; = access to technical expertise and public-interest compute; = sector-specific use cases and controlled testing environments; = measurement of benefits, failures and unintended consequences.

  4. Support Australian domain models where public value and lawful data use justify them.

    support Australian domain models models where public value and lawful data use justify them.

  5. Develop national data-sovereignty assurance focused on material data journeys, jurisdictions, subprocessors, model interactions, retention, portability and exit.

    Develop national data-sovereignty assurance focused on material data journeys, jurisdictions, subprocessors, model interactions, retention, portability and exit.

  6. Establish measurable public-interest compute access for Australian SMEs, researchers, not-for-profits and public-value applications, and assess provider concentration and continuity risk.

    Establish measurable public-interest compute access for Australian SMEs, researchers, not-for- profits and public-value applications, and assess provider concentration and continuity risk.

  7. Connect infrastructure investment to durable pathways through universities, TAFEs, schools and apprenticeships.

    Infrastructure investment should also connect to durable pathways through universities, TAFEs, schools and apprenticeships.

  8. Require proportionate AI registers and assurance documentation for material Commonwealth AI uses, including accountable ownership, evaluation, human oversight, incident learning and dependency visibility.

    Require proportionate AI registers and assurance documentation for material Commonwealth AI uses, including accountable ownership, evaluation, human oversight, incident learning and dependency visibility.

  9. For material Commonwealth AI systems, identify purpose, accountable owner, lawful basis, data sources, risk classification, evaluation, human oversight, incidents, significant vendors, dependencies and review arrangements.

    The Commonwealth can set a practical benchmark. For material AI systems, agencies should be able to identify purpose, accountable owner, lawful basis, data sources, risk classification, evaluation, human oversight, incidents, significant vendors, dependencies and review arrangements.

  10. For higher-impact Commonwealth AI uses, require a documented necessity and proportionality case, third-party impact assessment, written decision record, independent assurance and periodic review.

    For higher-impact Commonwealth AI uses, require a documented necessity and proportionality case, third-party impact assessment, written decision record, independent assurance and periodic review.

  11. For higher-impact Commonwealth AI, require accountable human oversight and written reasons alongside necessity/proportionality, third-party impact consideration, independent assurance and periodic review.

    For higher-impact Commonwealth AI uses, require a documented necessity and proportionality case, consideration of third-party impacts, accountable human oversight, written reasons, independent assurance and periodic review.

    A program corrected obvious text-recognition errors in this quote for display.

  12. Reassess AI systems when model, data, vendor, purpose or risk materially changes.

    Material changes to model, data, vendor, purpose or risk should trigger reassessment.

  13. Support clear provenance, licensing and transparency expectations for material use of Australian creative, cultural and media content in AI training and deployment, with proportionate mechanisms for rights-holder control and dispute resolution.

    Support clear provenance, licensing and transparency expectations for material use of Australian creative, cultural and media content in AI training and deployment, with proportionate mechanisms for rights-holder control and dispute resolution.

  14. Ensure safety frameworks particularly address children, older Australians, disability, low digital literacy and essential-service users; do not leave synthetic-content detection or automated-decision challenges solely to those least equipped.

    Safety frameworks should pay particular attention to children, older Australians, people with disability, people with lower digital literacy and people interacting with essential services. The burden of detecting synthetic content or challenging automated outcomes should not fall solely on the person least equipped to do so.

  15. Establish a cross-government AI dependency and resilience assessment for critical and high-impact uses, covering provider concentration, cyber risk, continuity, substitutability, foreign dependency and exit capability.

    Establish a cross-government AI dependency and resilience assessment for critical and high-impact uses, covering provider concentration, cyber risk, continuity, substitutability, foreign dependency and exit capability.

  16. Apply a National & Community Benefit Assessment to major AI/data-centre infrastructure, with enforceable Community Benefit Agreement pathways where projects substantially affect local resources or infrastructure.

    Apply an NCBA to major AI/data-centre infrastructure and create pathways for enforceable Community Benefit Agreements where projects have substantial local resource or infrastructure impacts.

    A program corrected obvious text-recognition errors in this quote for display.

  17. Assess major AI/data-centre projects’ enduring public value, potentially covering energy additionality and network costs, water sustainability, Australian skills and supply chains, environment, resilience, compute access and community benefit.

    Major AI/data-centre infrastructure should be assessed not only by whether a site can accommodate it, but by where the investment can create the greatest enduring public value. An NCBA could assess energy additionality and network costs; water sustainability; Australian skills and supply chains; environment and resilience; compute access; and local/community benefit.

  18. Where developments materially affect local resources and infrastructure, convert promises through Community Benefit Plans or Agreements into measurable long-term outcomes rather than discretionary sponsorship.

    Where developments materially affect local resources and infrastructure, Community Benefit Plans or Agreements should convert material promises into measurable long-term outcomes rather than discretionary sponsorship.

  19. Recommend national measures of safety/cyber/assurance/incidents, trust, productivity/services, infrastructure community benefits and domestic access, energy/water additionality and flexibility, retained research/skills/IP/engineering/supply chains, and adoption by SMEs, family businesses, not-for-profits and regions.

    The Committee should recommend a small set of national outcome measures so Australia can determine whether its AI strategy is delivering public value over time: = safety, cyber, assurance and incident outcomes; = public trust and confidence in significant AI use. = productivity and service improvements realised; = regional and community benefits created by AI infrastructure; = Australian community accessible to domestic users at a viable cost; = energy/water additionality and verified flexibility of major AI infrastructure; = Australian research, skills, IP, engineering and supply-chain capability retained; = AI adoption by SMEs, family businesses, not-for-profits and regional organisations;

    A program corrected obvious text-recognition errors in this quote for display.

Read the original

In some quotes a program corrected obvious text-recognition errors for display.

Open in the interactive guide