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 390

Mrs Elizabeth Morrish

Submitter group: Individuals.

Position labels are not shown for individual submitters.

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.

An individual argues for a precautionary, risk-based approach with mandatory safeguards on frontier AI while avoiding unnecessary burdens on lower-risk adoption, and says AI has enormous potential. Covers independent safety evaluation, the AI Safety Institute, sovereign capability, distributional effects, international standards and transparency about environmental impacts. The author acknowledges uncertainty about catastrophic-risk likelihood.

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 frontier developers should demonstrate adequate mitigation of foreseeable catastrophic risks to independent regulators.

    For frontier systems, society should not have to prove that a technology is dangerous after it has been developed or deployed. Developers need to demonstrate to independent regulators that foreseeable catastrophic risks have been adequately mitigated.

  2. Argues that companies should not decide on society's behalf what level of severe or irreversible risk is acceptable.

    Where a technology creates potentially severe or irreversible risks, the companies developing it should not be able to determine on society's behalf what level of risk is acceptable.

  3. Argues that Australian businesses adopting lower-risk AI should not face unnecessary barriers, and that the strongest requirements can focus on frontier developers and operators.

    Australian businesses adopting lower-risk AI to improve productivity do not need unnecessary barriers simply because they use AI. The strongest requirements can instead focus on developers and operators of frontier systems capable of creating the greatest harm.

  4. Argues that AI benefits may not be broadly distributed and could increase inequality if returns accrue to owners of capital.

    AI could create enormous economic value, but there is no guarantee that those benefits will be broadly distributed. If AI increasingly substitutes for human labour while returns accrue predominantly to the owners of AI systems, computing infrastructure and capital, it could contribute to greater economic inequality.

  5. Argues that international coordination on common minimum standards matters so that investment and risk do not simply migrate.

    If responsible jurisdictions impose meaningful safety requirements while others allow frontier development without equivalent safeguards, both investment and risk may migrate rather than the underlying problem being solved. Australia has an interest in pursuing common minimum standards among countries developing advanced AI.

  6. Argues that Australians need transparent information about the environmental footprint of large-scale AI.

    Australians need transparent information about the environmental footprint of large-scale AI development and deployment.

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 proportionate, risk-based AI regulation, avoiding unnecessary burdens for lower-risk commercial adoption while applying mandatory safeguards to frontier systems.

    Establish proportionate, risk-based AI regulation, avoiding unnecessary compliance burdens for lower-risk commercial adoption while applying mandatory safeguards to frontier systems.

  2. Give the AI Safety Institute sufficient independence, expertise, funding and access to advanced models.

    Give the Australian AI Safety Institute sufficient independence, expertise, funding and access to advanced models to conduct meaningful safety evaluations.

  3. Require independent pre-deployment testing and serious-incident reporting for frontier systems, with regulatory powers to delay or prevent deployment where catastrophic risks are inadequately controlled.

    Require independent pre-deployment testing and serious-incident reporting for frontier systems, with regulatory powers to delay or prevent deployment where catastrophic risks have not been adequately controlled.

  4. Pursue international minimum standards for frontier AI safety.

    Pursue international minimum standards for frontier AI safety so that responsible regulation does not simply shift investment and risk to less regulated jurisdictions.

  5. Support Australian AI skills, research, secure computing infrastructure and hosting capability to strengthen economic competitiveness, data sovereignty and national resilience.

    Support Australian AI skills, research, secure computing infrastructure and hosting capability to strengthen economic competitiveness, data sovereignty and national resilience.

  6. Monitor entry-level and youth employment, career pathways, real wages, labour’s share of national income, concentration of economic power and the distribution of productivity gains.

    Government monitoring of entry-level employment and career pathways, youth employment, real wages, labour's share of national income, concentration of economic power and the distribution of productivity gains will be important.

  7. Respond to AI’s distributional effects on entry-level opportunities, employment, wages, inequality and the sharing of productivity gains.

    Monitor and respond to the distributional effects of AI on entry-level opportunities, employment, wages, inequality and the sharing of productivity gains.

  8. Ensure Australia's AI governance approach considers intergenerational impacts and supports meaningful representation of developing nations in international AI governance.

    Ensure Australia’s approach to AI governance considers intergenerational impacts and supports meaningful representation of developing nations in international AI governance.

  9. Require transparency about the energy, water and environmental impacts of large-scale AI infrastructure.

    Require meaningful transparency about the energy, water and environmental impacts of large-scale AI infrastructure.

  10. Have education for an AI-enabled economy include AI literacy while preserving independent thought, creativity, literacy, numeracy, interpersonal skills and the capacity to question automated outputs.

    Education for an AI- enabled economy needs to include AI literacy while preserving independent thought, creativity, literacy, numeracy, interpersonal skills and the capacity to question automated outputs.

  11. Require frontier developers to demonstrate adequate mitigation of foreseeable catastrophic risks to independent regulators, with mandatory independent evaluation and clearly defined capability thresholds for stronger requirements.

    For frontier systems, society should not have to prove that a technology is dangerous after it has been developed or deployed. Developers need to demonstrate to independent regulators that foreseeable catastrophic risks have been adequately mitigated. This requires mandatory independent safety evaluation, clearly defined capability thresholds for stronger requirements, and regulatory powers to delay or prevent deployment where adequate safety cannot be demonstrated.

  12. Include environmental impacts in Australia’s overall AI strategy and prevent environmental costs being externalised to households or communities while returns are privatised.

    The environmental cost should not be externalised to households or communities while economic returns are privatised, and these impacts need to form part of Australia's overall AI strategy.

Read the original

Open in the interactive guide