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 188

National Farmers' Federation

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.

Proposes evidence-based, proportionate and technology-neutral AI policy, balancing productivity and regional opportunities with risks. Recommends mapping existing law before new obligations, operational definitions of sovereignty, worker consultation and adjustment, safeguards for consequential government AI, and coordinated, reviewable policy. Also asks the Committee to examine copyright, child safety, security and environmental impacts.

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. Recommends that the Committee apply a risk-based, technology-neutral and outcomes-focused approach, allowing targeted obligations for some uses.

    The Committee should apply a risk-based, technology-neutral and outcomes-focused approach, while recognising that some AI uses may justify targeted obligations.

  2. Proposes the principle that new measures should address an identified deficiency and add value beyond existing laws, regulators and standards.

    Evidence and additionality: new measures should address an identified deficiency and add value beyond existing laws, regulators and standards.

  3. Says a dedicated AI Act should neither be assumed nor excluded at the outset, and any case for new legislation should set out its objective, scope and costs.

    A dedicated AI Act should not be assumed or excluded at the outset. Any case for new legislation should identify the regulatory objective, scope, responsible parties, interaction with existing law, compliance costs, enforcement arrangements and review mechanisms.

  4. States that full domestic self-sufficiency may be neither feasible nor necessary and that the objective should be the level of capability and control needed for resilience, security and economic opportunity.

    Full domestic self-sufficiency may be neither feasible nor necessary; the policy objective should be the level of capability and control required for resilience, security and economic opportunity.

  5. Says the Committee should seek disaggregated evidence on workforce effects and not treat AI exposure as equivalent to job displacement.

    NFF believes that the Committee should seek disaggregated evidence and avoid treating exposure to AI as equivalent to job displacement.

  6. Recommends higher-impact Commonwealth AI safeguards including a legally accountable decision-maker, meaningful human oversight, audit records, appropriate notice, review pathways, procurement assurance and independent evaluation.

    Commonwealth policy for responsible use of AI, including accountable officials, transparency statements, internal registers, staff training, impact assessment and incident management. For higher-impact uses, minimum safeguards should include a legally accountable decision-maker, meaningful human oversight, records sufficient for audit, clear notice where appropriate, accessible review pathways, procurement assurance and independent evaluation.

Recommendations

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

  1. Adopt a risk-based, technology-neutral and outcomes-focused framework for assessing all terms of reference.

    Adopt a risk-based, technology-neutral and outcomes-focused framework for assessing all terms of reference.

  2. Map existing AI-relevant laws and assess clarity, regulator powers and technical capability, and enforcement/remedy gaps.

    The Committee should map the application of existing privacy, consumer, competition, corporations, employment, work health and safety, discrimination, administrative, criminal and sector-specific laws. The central questions are whether those laws apply clearly, whether regulators have adequate powers and technical capability, and whether enforcement or remedy gaps arise in practice.

  3. Identify regulatory gaps by reference to evidence of harm, uncertainty, inadequate remedies or enforcement limitations.

    Identify regulatory gaps by reference to evidence of harm, uncertainty, inadequate remedies or enforcement limitations.

  4. Assess the costs and benefits of options for small and medium enterprises, family businesses, regional communities and not-for-profit organisations.

    Assess the costs and benefits of options for small and medium enterprises, family businesses, regional communities and not-for-profit organisations.

  5. Distinguish general-purpose and domain-specific AI uses.

    Seek sector-specific evidence, including from agriculture and other areas of Australian comparative advantage, and distinguish between general-purpose and domain-specific uses.

  6. Recommend practical skills, management capability, safe adoption and workforce adjustment measures through consultation with employers, workers and education providers.

    Recommend practical measures to improve skills, management capability, safe adoption and workforce adjustment, supported by consultation with employers, workers and education providers.

  7. For higher-impact Commonwealth AI, require accountable decision-makers, meaningful human oversight, audit records, appropriate notice, accessible review, procurement assurance and independent evaluation; balance transparency against privacy, security and legitimate confidentiality.

    Government also bears heightened responsibilities because its decisions can affect rights, entitlements and access to essential services. The Committee should examine implementation of the Commonwealth policy for responsible use of AI, including accountable officials, transparency statements, internal registers, staff training, impact assessment and incident management. For higher-impact uses, minimum safeguards should include a legally accountable decision-maker, meaningful human oversight, records sufficient for audit, clear notice where appropriate, accessible review pathways, procurement assurance and independent evaluation. Transparency should be balanced against privacy, security and legitimate confidentiality.

  8. Define sovereign capability and data sovereignty in operational terms, linked to identified resilience, security and economic objectives.

    Define sovereign capability and data sovereignty in operational terms, linked to identified resilience, security and economic objectives.

  9. Promote national consistency and international interoperability while avoiding unnecessary duplication across jurisdictions and regulators.

    Promote national consistency and international interoperability while avoiding unnecessary duplication across jurisdictions and regulators.

  10. Provide staged recommendations: immediate administrative actions, medium-term regulatory reforms and matters requiring further research or consultation.

    Provide staged recommendations: immediate administrative actions, medium-term regulatory reforms and matters requiring further research or consultation.

  11. Include review and evaluation mechanisms so policy settings can respond to technological and market developments.

    Include review and evaluation mechanisms so that policy settings can respond to technological and market developments.

  12. Calibrate policy to harm likelihood and severity, use context and affected entities' capacity.

    Proportionality: policy responses should be calibrated to the likelihood and severity of harm, the context of use and the capacity of affected entities.

  13. Focus obligations on functions, impacts and outcomes, allowing targeted treatment where distinctive risks are demonstrated.

    Technology neutrality: obligations should generally focus on functions, impacts and outcomes rather than particular models or products, while allowing targeted treatment where distinctive risks are demonstrated.

  14. Clarify lifecycle responsibilities for developers, deployers, procurers and users according to practical control over risk.

    Accountability across the lifecycle: responsibilities should be clear for developers, deployers, procurers and users, having regard to their practical control over risks.

  15. Make consequential decisions explainable, reviewable and capable of appropriate human intervention.

    Contestability and human oversight: consequential decisions should be explainable, reviewable and capable of appropriate human intervention.

  16. Examine when AI produces measurable economic and social value and how gains are distributed, rather than treating adoption as an end.

    NFF would submit that the Committee should examine the conditions under which AI produces measurable economic and social value, rather than treating adoption as an end in itself. Relevant considerations include productivity, service quality, resilience, diffusion of innovation and the distribution of gains between firms, workers, consumers and communities.

  17. Distinguish adoption barriers warranting public intervention from ordinary commercial constraints; consider guidance, shared assurance, demonstrations, standards support, training and independent advice.

    The Committee, in NFF’s view, should distinguish between barriers that warrant public intervention and ordinary commercial constraints. Options for consideration include practical guidance, shared assurance tools, demonstration projects, standards support, workforce training and improved access to independent advice.

  18. Account for privacy, confidentiality, Indigenous data governance, IP, data quality and cyber security in domain models; consider trusted access and data sharing preserving legitimate rights and safeguards.

    Domain-specific models trained on Australian data may offer advantages in sectors such as agriculture, health and public administration, but their development must account for privacy, confidentiality, Indigenous data governance, intellectual property, data quality and cyber security. The Committee should consider models for trusted access and data sharing that preserve legitimate rights and safeguards.

  19. Seek disaggregated workforce evidence and avoid equating AI exposure with job displacement.

    NFF believes that the Committee should seek disaggregated evidence and avoid treating exposure to AI as equivalent to job displacement.

  20. Support fair workforce adjustment without assuming uniformly positive or negative employment effects.

    Workforce policy should support consultation, skills development, safe deployment and fair adjustment, without assuming either uniformly positive or uniformly negative employment effects.

  21. Assess existing workplace laws and institutions before adding obligations; provide timely targeted adjustment support where change is rapid or concentrated.

    Existing workplace laws and institutions should be assessed before recommending additional obligations. Where change is rapid or concentrated, adjustment support should be timely and targeted.

  22. Examine implementation of Commonwealth responsible-AI policy, including accountable officials, transparency statements, internal registers, staff training, impact assessment and incident management.

    The Committee should examine implementation of the Commonwealth policy for responsible use of AI, including accountable officials, transparency statements, internal registers, staff training, impact assessment and incident management.

  23. Publish the existing-law, regulator, standards and initiatives map before recommending new obligations.

    Publish a clear map of existing laws, regulators, standards and government initiatives relevant to AI before recommending new obligations.

  24. Neither assume nor exclude an AI Act; require any new-legislation case to identify objectives, scope, responsible parties, existing-law interactions, costs, enforcement and review.

    A dedicated AI Act should not be assumed or excluded at the outset. Any case for new legislation should identify the regulatory objective, scope, responsible parties, interaction with existing law, compliance costs, enforcement arrangements and review mechanisms.

  25. Consider creators', cultural institutions', media, researchers', developers' and users' interests in copyright; assess training access, transparency, licensing, attribution, remuneration, research/innovation exceptions, enforcement and outputs, consistently with international obligations and market evidence.

    The inquiry should consider the legitimate interests of creators, cultural institutions, media organisations, researchers, technology developers and users. Key issues include lawful access to training material, transparency, licensing, attribution, remuneration, exceptions for research and innovation, enforcement practicality and the treatment of outputs. Recommendations should be consistent with Australia’s international obligations and supported by evidence about market impacts.

  26. Assess each AI risk by affected persons, causal pathways, scale, reversibility and adequacy of prevention, redress and enforcement.

    The Committee should assess each risk by reference to affected persons, causal pathways, scale, reversibility and the adequacy of existing prevention, redress and enforcement mechanisms.

  27. Evaluate provenance, proportionate provider/platform duties, awareness, scam disruption, age safeguards, testing, monitoring, incident reporting and remedies for effectiveness, enforceability and unintended effects on expression, research and innovation.

    Potential measures include authentication and provenance standards, platform and provider duties proportionate to control, public awareness, stronger scam disruption, age-appropriate safeguards, testing and monitoring, incident reporting and accessible remedies. Measures should be evaluated for effectiveness, enforceability and unintended effects on legitimate expression, research and innovation.

  28. Examine secure procurement, supply-chain assurance, model/data security, evaluation capability, workforce expertise and government–industry information sharing.

    The Committee should examine secure procurement, supply-chain assurance, model and data security, evaluation capability, workforce expertise and information sharing between government and industry.

  29. Preserve lawful oversight, accountability and civil liberties in national-security measures; distinguish commercial AI, dual-use capabilities and systems directly affecting defence, intelligence or critical infrastructure.

    National security measures should preserve lawful oversight, accountability and civil liberties. Recommendations should distinguish between ordinary commercial AI, dual-use capabilities and systems with direct implications for defence, intelligence or critical infrastructure.

  30. Seek trusted international interoperability where it reduces costs and supports trade, retaining capacity to address Australian legal, cultural and strategic priorities.

    Australia should seek interoperability with trusted international standards and assurance practices where this reduces compliance costs and supports trade, while retaining the ability to address Australian legal, cultural and strategic priorities.

  31. Consider AI computing's energy, water, land-use and infrastructure impacts and efficiency opportunities, with proportionate reporting/policy responses coordinated across jurisdictions.

    The Committee should also consider the energy, water, land-use and infrastructure implications of AI computing, together with opportunities for efficiency and innovation. Reporting and policy responses should be proportionate to material impacts and coordinated across jurisdictions

  32. Integrate economic, social, legal, workforce, national-security and public-sector considerations in the inquiry.

    The inquiry should integrate economic, social, legal, workforce, national security and public- sector considerations.

  33. Seek sector-specific evidence from agriculture, resources, health/medical research, advanced manufacturing, financial services and defence, avoiding assumptions that one adoption model suits all sectors or locations.

    Sector-specific evidence should be sought from agriculture, resources, health and medical research, advanced manufacturing, financial services and defence industry, while avoiding assumptions that a single model of adoption will suit all sectors or locations.

Read the original

Open in the interactive guide