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 290

Business Council of Australia (BCA)

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.

Open in the interactive guide

Summary

Read the original submission for the full text.

Argues for business-led AI adoption with trusted global access and sovereign control of critical functions. Prefers existing laws and targeted gap-filling, voluntary copyright licensing and lawful training access unless creators opt out, workforce and R&D capability, flexible national data-centre standards, and evidence-based public safeguards.

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 AI can help businesses and workers do more with available resources, improve service quality and speed, and support new products, firms and industries.

    AI can help respond to these challenges by enabling businesses and workers to do more with the resources available, improve the quality and speed of services, and support new products, firms and industries.

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

  2. Argues that sovereign AI does not require building every capability domestically, but ensuring critical national functions can continue during disruption.

    The objective is not to build every AI capability in Australia but to ensure critical national functions can continue to operate in periods of heightened geopolitical tension, supply-chain disruption, or restricted access to overseas technology.

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

  3. Argues that government should identify a material gap and explain why existing tools cannot address it before introducing new AI-specific obligations.

    Before introducing any new AI-specific obligation, Government should identify the material harm or accountability gap, explain why existing law, regulator guidance, standards or business practice cannot address it, and assess the costs and interactions of the proposed response.

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

  4. Proposes a clear legal basis for AI training on lawfully accessed material unless owners opt out, preserving voluntary licensing after an opt-out.

    The BCA’ preferred approach is that AI developers should have a clear legal basis to use lawfully accessed material for AI model training unless the copyright owner has signalled that the material should not be used. Where a rights holder opts out, the developer and rights holder should remain free to enter a voluntary licensing arrangement.

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

  5. Argues that government should curate trusted foundational AI learning material underpinned by an industry-informed capability framework.

    Government should make trusted foundational AI learning easier to access by curating high-quality training material from government, industry and training providers, underpinned by an industry-informed capability framework.

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

  6. Recommends aligning Commonwealth, state and territory data centre settings in a single national framework.

    Align Commonwealth, state and territory settings for data centres and their supporting energy, water and connectivity infrastructure through a single national framework.

Recommendations

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

  1. Secure enduring access to leading global AI technology and trusted international suppliers.

    Secure enduring access to leading global AI technology and trusted international suppliers.

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

  2. Build and leverage Australia's strengths in energy, critical minerals, sector expertise, data, research, infrastructure and trusted institutions.

    Build and leverage Australia’s strengths in energy, critical minerals, sector expertise, data, research, infrastructure and trusted institutions.

  3. Retain sufficient domestic capability and control to sustain critical national functions, including domestically controlled model capability where the consequences of external dependence would be unacceptable.

    Retain sufficient domestic capability and control to sustain critical national functions, including domestically controlled model capability where the consequences of external dependence would be unacceptable.

  4. Coordinate AI-related reforms, align definitions and implementation, and avoid overlapping or duplicate obligations.

    Coordinate AI-related reforms, align definitions and implementation, and avoid overlapping or duplicate obligations.

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

  5. Use existing technology-neutral laws wherever they address the risk.

    Use existing technology-neutral laws wherever they address the risk.

  6. Limit AI-specific obligations to clear and material gaps, supported by proportionate impact analysis.

    AI-specific obligations should be limited to clear and material gaps and supported by proportionate impact analysis.

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

  7. Maintain voluntary licensing as the backbone of Australia's copyright framework and preserve copyright owners' autonomy to enter commercial licensing arrangements.

    Maintain voluntary licensing as the backbone of Australia’s copyright framework and preserve copyright owners’ autonomy to enter commercial licensing arrangements.

  8. Support and honour simple, low-burden and internationally recognised mechanisms allowing copyright owners to opt out of AI model training.

    Support and honour simple, low-burden and internationally recognised mechanisms that allow copyright owners to opt out of AI model training.

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

  9. Provide greater legal certainty about how the Copyright Act applies to the use of lawfully accessed material for AI model training.

    Provide greater legal certainty about how the Copyright Act applies to the use of lawfully accessed material for AI model training.

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

  10. Undertake further consultation before giving opt-out signals greater legal effect, and avoid compulsory licensing or embedded payment obligations.

    Undertake further consultation before giving opt-out signals greater legal effect and avoid compulsory licensing or embedded payment obligations.

  11. Recognise that existing workplace relations, work health and safety, privacy, surveillance and anti-discrimination laws contain adequate safeguards for workplace AI, including where AI informs significant employment decisions.

    Recognise that existing workplace relations, work health and safety, privacy, surveillance and anti- discrimination laws contain adequate and appropriate safeguards for workplace AI, including where AI informs decisions with significant employment consequences.

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

  12. Preserve employers' capacity to redesign work and realise productivity gains within existing consultation, safety and legal obligations.

    Preserve employers’ capacity to redesign work and realise productivity gains within existing consultation, safety and legal obligations.

  13. Maintain clear human accountability, transparency and explainability on the data used in decisions where AI contributes materially to significant employment outcomes.

    Maintain clear human accountability, transparency and explainability on the data used in decisions where AI contributes materially to significant employment outcomes.

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

  14. Support workforce engagement and responsible implementation without creating a general veto over technology or work design.

    Support workforce engagement and responsible implementation without creating a general veto over technology or work design.

  15. Maintain employer responsibility for role-specific and workplace-based AI training, including the practical application of AI to business processes, data and risk.

    Maintain employer responsibility for role-specific and workplace-based AI training, including the practical application of AI to business processes, data and risk.

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

  16. Curate high-quality foundational AI learning under an industry-informed capability framework, covering practical use, verification, security, privacy and responsible application.

    Government should make trusted foundational AI learning easier to access by curating high-quality training material from government, industry and training providers, underpinned by an industry-informed capability framework. The training suite should cover practical use, verification, security, privacy and responsible application.

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

  17. Target additional support to SMEs, jobseekers, career changers and regional and lower-access communities less likely to receive structured workplace training.

    Target additional support to small and medium-sized enterprises, jobseekers, career changers and regional and lower-access communities that are less likely to receive structured workplace training.

  18. Provide timely migration and intra-company transfer pathways for genuinely scarce AI and frontier-technology specialists, complementing rather than substituting for investment in Australian workers and researchers.

    Provide timely migration and intra-company transfer pathways for genuinely scarce AI and frontier- technology specialists, complementing rather than substituting for investment in Australian workers and researchers.

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

  19. Implement Strategic Examination of R&D reforms, including National Innovation Pillars and National Strategic Initiatives, focusing AI research and commercialisation on comparative advantage, distinctive data or strategic need.

    Implement the Strategic Examination of R&D reforms to translate Australia’s AI strengths into commercial outcomes, including the National Innovation Pillars and National Strategic Initiatives, focusing AI research and commercialisation on areas where Australia has comparative advantage, distinctive data or strategic need.

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

  20. Improve compute access for innovative Australian startups, small businesses, researchers and nonprofits where materially constrained, preferring transparent commercial mechanisms over mandatory set-asides, with clear eligibility, security, evaluation and additional-outcome requirements.

    The BCA supports improving access for innovative Australian start-ups and small businesses, researchers and not-for-profits where compute is a material constraint on research, development or commercialisation. Commercial mechanisms such as transparent provider partnerships, competitive credits, co- funded programs and shared research infrastructure should be preferred to mandatory capacity set-asides, with clear eligibility, security and evaluation arrangements and a focus on additional research and commercialisation outcomes.

  21. Align Commonwealth, state and territory settings for data centres and their supporting energy, water and connectivity infrastructure through a single national framework.

    Align Commonwealth, state and territory settings for data centres and their supporting energy, water and connectivity infrastructure through a single national framework.

  22. Make the framework national and proportionate: lighter requirements for smaller facilities, full standard for large new or expanded facilities above a materiality threshold.

    use one Commonwealth-state framework, with lighter requirements for smaller facilities and the full standard reserved for large new or expanded facilities above a defined materiality threshold

  23. Base sustainability reporting on metered, post-operation energy, water and emissions data, with proportionate assurance, rather than planned or projected usage.

    measured and verified: base sustainability reporting on metered, post-operation energy, water and emissions data, supported by proportionate assurance, rather than planned or projected usage

  24. Require connecting data centres to pay the efficient, incremental and independently attributable costs of connection and upstream network augmentation they trigger, subject to regulatory oversight, while enabling private investment in shared assets.

    require connecting data centres to pay the efficient, incremental and independently attributable costs of their connection and any upstream network augmentation they trigger, subject to regulatory oversight, while enabling private investment in shared assets

  25. Publish information at portfolio or aggregate level, while providing site-level data securely and confidentially to regulators, utilities and government where required.

    publish information at portfolio or aggregate level, while site-level data is provided securely and confidentially to regulators, utilities and government where required

  26. Give data centres that meet the standard expedited decisions, coordinated infrastructure planning and regulatory certainty, and avoid moratoria, capacity caps and retrospective rules.

    give data centres that meet the standard expedited decisions, coordinated infrastructure planning and regulatory certainty, and avoid moratoria, capacity caps and retrospective rules.

  27. Preserve flexible, efficient, commercially workable infrastructure outcomes, avoiding fragmented, prescriptive or retrospective rules; support new renewable generation alongside renewable energy, storage and firming, including gas where required.

    Preserve flexibility in how projects meet the required outcomes and avoid fragmented, prescriptive or retrospective requirements. This includes energy settings that support new renewable generation while preserving flexibility for data centres to use an efficient and commercially workable mix of renewable energy, storage and firming, including gas where required.

  28. Build a credible public evidence base on AI benefits and impacts, drawing on existing data, evaluated case studies and aggregate sector-level information.

    Build a credible public evidence base on the benefits and impacts of AI across the economy, drawing on existing data, evaluated case studies and aggregate sector-level information.

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

  29. Have government lead responsible AI deployment and publicly report measurable outcomes, within an evidence base including industry case studies and protections for privacy, security and commercial confidentiality.

    Government and industry should lead by example through responsible AI deployment and credible evidence of measurable benefits, including: - public reporting by government agencies on the outcomes of AI uses - industry-led evidence and practical case studies demonstrating improvements in productivity, safety, service quality and customer outcomes - appropriate protection for privacy, security and commercial confidentiality.

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

  30. Use the evidence to target enforcement and any further reform at demonstrated harms, avoiding broad constraints on low-risk and beneficial AI uses.

    Use that evidence to target enforcement and any further reform at demonstrated harms, while avoiding broad constraints on low-risk and beneficial AI uses.

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

  31. Keep markets open, international partnerships and investment contestable, avoiding local-content requirements protecting uncompetitive inputs; support value capture through stable competitive policy, timely infrastructure, skills, research and procurement demonstration opportunities.

    Government can enable this outcome through stable and competitive policy settings, timely infrastructure, access to skills and research, and procurement that allows new solutions to demonstrate value. Open markets, international partnerships and contestable investment should remain central; local-content requirements that protect uncompetitive inputs would weaken the ecosystem.

  32. For the most critical national functions, consider domestic operation, testing and adaptation of large language models, including frontier capability particularly for national-security and cybersecurity functions.

    For the most critical national functions, this may require large language models and the ability to operate, test and adapt those models in Australia. This should include frontier capability, particularly for national security functions including cyber security. Australia needs the best AI models to defend against rapidly evolving Al- enabled threats.

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

  33. Concentrate government AI investment and intervention on critical national functions, risks and market failures.

    Government investment and intervention should be concentrated on critical national functions, risks and market failures.

  34. Before introducing any new AI-specific obligation, identify the material harm or accountability gap, explain why existing law, guidance, standards or business practice cannot address it, and assess costs and interactions.

    Before introducing any new AI-specific obligation, Government should identify the material harm or accountability gap, explain why existing law, regulator guidance, standards or business practice cannot address it, and assess the costs and interactions of the proposed response.

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

  35. Make Australian AI Standards principles-based, risk-proportionate, internationally interoperable and complementary to existing law; avoid prescriptive technical requirements that may become outdated or undermine effectiveness.

    The Government's intention to establish nationally consistent Australian Standards for AI'* provides an opportunity to give businesses practical guardrails for responsible adoption, strengthen public confidence and provide greater certainty for investment and deployment. The standards should be principles-based, proportionate to risk, interoperable with relevant international standards and designed to complement rather than duplicate existing legal obligations. They should avoid overly prescriptive technical requirements that could quickly become outdated or undermine their effectiveness

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

  36. Include realistic transition periods, clear regulator responsibility and scheduled review in new AI-specific obligations.

    Before introducing any new AI-specific obligation, Government should identify the material harm or accountability gap, explain why existing law, regulator guidance, standards or business practice cannot address it, and assess the costs and interactions of the proposed response. Measures should include realistic transition periods, clear regulator responsibility and scheduled review, consistent with regulatory best practice.

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

  37. Keep statutory regulators accountable within existing mandates, supported by consistent guidance and technical bodies' evidence, testing and horizon scanning.

    Technical bodies can support regulators with evidence, testing and horizon scanning. Statutory regulators should remain accountable for applying the law within their existing mandates, supported by consistent guidance.

  38. Make transparency sufficient for accountability, material-use understanding and oversight, while protecting cybersecurity, privacy, IP and commercial information; make reporting risk-based and reusable where regulators request the same evidence.

    Information should be sufficient to identify accountability, understand material uses and enable oversight, but requirements should protect cyber security, privacy, intellectual property and commercially sensitive information. Reporting should be risk-based and reusable across regulators where the same evidence is requested.

  39. Give AI developers a clear legal basis for training on lawfully accessed material unless owners opt out; preserve voluntary licensing after opt-outs.

    The BCA’ preferred approach is that AI developers should have a clear legal basis to use lawfully accessed material for AI model training unless the copyright owner has signalled that the material should not be used. Where a rights holder opts out, the developer and rights holder should remain free to enter a voluntary licensing arrangement.

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

  40. Apply ordinary copyright principles to AI, protecting expression while keeping underlying ideas, themes, concepts and methods available for use and development.

    Copyright law should be applied to AI in the same way it applies to everything else. This protects expression while keeping the underlying ideas, themes, concepts, and methods free for everyone to use and build upon.

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

  41. Have the AI Safety Institute assess emerging capabilities, monitor risks and harms, and provide technical expertise to regulators, strengthening evidence for effective coordinated responses.

    The work of Australia’s AI Safety Institute should support this approach by assessing emerging AI capabilities, monitoring risks and harms, and providing technical expertise to help regulators respond. Its work should strengthen the evidence base for identifying risks and developing effective, coordinated responses.

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

  42. Consider additional AI-specific workplace regulation only on clear evidence of a material gap unaddressable by existing laws, guidance, implementation support, industry standards and good business practice.

    Additional AI-specific workplace regulation should only be considered in future if there is clear evidence of a material gap that cannot be addressed through the comprehensive existing laws, regulator guidance, practical implementation support, industry standards and good business practice. The BCA has not identified any such gaps at this time.

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

  43. Sequence reforms methodically, assessing existing/proposed risk mitigation and avoiding duplication; give businesses reasonable time and clear guidance to implement compliance.

    A methodical approach should assess how existing and proposed reforms mitigate identified risks, avoid duplication and ensure coherence across regulatory regimes. Businesses need reasonable time and clear guidance to familiarise themselves with applicable requirements and implement the systems and processes needed for compliance.

  44. Connect foundational AI learning to specialised employer-led, industry-led and accredited training.

    Foundational learning should connect to progressively more specialised employer-led, industry-led and accredited training.

  45. Evaluate training programs by adoption and capability outcomes, beyond participation numbers.

    Programs should be evaluated against adoption and capability outcomes, not participation numbers alone.

  46. Leverage Australia's medical research and clinical-trials strengths to attract investment in AI-enabled diagnosis, treatments and medicines.

    One of the areas where AI provides a significant opportunity in the development of new treatments for health conditions, through early diagnosis to new medicines. Australia’s strength in medical research, including clinical trials, should be leveraged to attract more investment in this area, delivering lifesaving benefits to Australians.

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

  47. Respond to SERD as an integrated reform agenda; prioritise mission-oriented industry/research collaboration, modern infrastructure, business-collaboration incentives and pathways from publicly funded research to deployment.

    The Government should respond to the SERD as an integrated reform agenda rather than through isolated programs. For AI, priorities should include mission-oriented industry-research collaboration in areas of comparative advantage, access to modern research infrastructure, stronger incentives for business collaboration, and clearer pathways from publicly funded research to commercial deployment.

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

  48. Keep procurement contestable, technology-neutral and value-for-money based, removing unnecessary barriers preventing credible Australian AI firms and partnerships from demonstrating capabilities.

    Government and large enterprise procurement can strengthen Australian AI R&D by providing early customers for new technologies, helping firms test, validate and scale products in real-world settings. Procurement should remain contestable, technology-neutral and based on value for money, while removing unnecessary barriers that prevent credible Australian firms and partnerships from demonstrating their capabilities.

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

  49. Design scale-up finance programs to crowd in private investment, support commercial milestones and financial independence, and develop products, management capability and international access while retaining meaningful Australian activity and IP.

    Access to scale-up finance is critical to turning promising AI R&D into commercial businesses. Programs should be designed to crowd in private investment, support commercial milestones and provide a pathway to financial independence. The objective is to help viable Australian firms build products, management capability and international market access while retaining meaningful activity and intellectual property in Australia.

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

  50. Do not preclude other compute models where strategically justified; allow government support for nationally governed research compute complementing commercial offerings and developing domestic technical skills.

    These mechanisms should not preclude other models where there is a clear strategic rationale. Governments may also choose to invest in or support nationally governed research compute capability, including facilities that complement commercial offerings and build domestic technical skills.

  51. Encourage efficient data-centre cooling and recycled or non-potable water supplies where feasible.

    Connecting data centres should meet the efficient, incremental and independently attributable costs of their connection and any upstream network augmentation they trigger, while water settings should encourage efficient cooling technologies and recycled or non-potable supply where feasible.

  52. Align data-centre principles, definitions, information requirements and assessment processes nationally, allowing genuine climate, grid and geographic differences; recognise existing planning, environment, energy, water and security rules and avoid duplication.

    Consistency should therefore extend to the principles, definitions, information requirements and assessment processes that apply to data-centre development, while allowing implementation to reflect genuine differences in climate, grid conditions and geography. Policy should also recognise the planning, environmental, energy, water and security frameworks that already apply and avoid unnecessary duplication.

  53. Have Commonwealth agencies publish meaningful information about material AI use, apply proportionate impact assessment and assurance, and maintain accountable human oversight.

    Commonwealth agencies should publish meaningful information about material AI use, apply impact assessment and assurance proportionate to the decision or service, and maintain accountable human oversight.

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

  54. Evaluate public-sector AI adoption against service quality, accessibility, productivity and trust.

    Public-sector adoption should be evaluated against service quality, accessibility, productivity and trust.

  55. Acknowledge uncertain AI outcomes in policy and update policy as evidence develops.

    Where outcomes are uncertain, policy should acknowledge uncertainty and update as evidence develops.

  56. Equip existing regulators to apply current laws against AI harms, with targeted reforms for demonstrated gaps.

    Harms including scams, fraud, deceptive synthetic content, unsafe products and abuse directed at children or vulnerable people require focused action. Existing regulators should be equipped to apply current laws, and targeted reforms should close demonstrated gaps.

  57. Account for unequal digital capability in AI policy, particularly in regional, remote and lower-income communities.

    AI policy should not assume that every person or business begins with the same digital capability, particularly in regional, remote and lower-income communities.

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

  58. Track trust and adoption through a common framework covering adoption/depth, workforce capability, public-sector outcomes, complaints/incidents, research/commercialisation results, infrastructure delivery and community confidence.

    Trust and adoption should be tracked through a common evidence framework. Governments should track measures including business adoption and depth of integration, workforce capability, public-sector outcomes, complaints and incidents, research and commercialisation results, infrastructure delivery and community confidence.

Read the original

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

Open in the interactive guide