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 421

Miss Kate Landry

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.

In a personal capacity, asks for independent evidence on catastrophic AI risk, verified international pacing agreements, a powerful AI Safety Institute, incident reporting, adaptable legislation and workforce scenarios. The author cites developers’ risk assessments rather than offering technical expertise, and describes both benefits so far in their own work and concern about future employability.

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 the submitter's central concern is that AI systems are improving faster than the ability to understand, verify and control them.

    My central concern is that these systems are improving faster than our ability to understand, verify and control them.

  2. Argues that a country cannot base its risk assessment on voluntary disclosures from companies answering to another jurisdiction.

    But a country cannot run its risk assessment on the voluntary disclosures of companies that answer to another jurisdiction. Australia needs to be able to find out for itself.

  3. Recommends AI Safety Institute powers to access frontier systems deployed in Australia, compel developer and deployer information, and give assessments consequences.

    Empower the Australian AI Safety Institute (item m) to do its job. Give it the power to require access to frontier systems deployed in Australia, to compel information from developers and deployers, and to have its assessments carry consequences.

  4. Recommends a reporting regime for serious AI incidents, by analogy with aviation incident reporting.

    Mandatory incident reporting (items i, l). A reporting regime for serious AI incidents. We require aviation incident reports so that near-misses are visible before a crash, not after.

  5. Describes concern about the future of the submitter's own role as AI automates more tasks.

    So what is left of my job in five years? Checking the emails it cannot handle? Agreeing with how it has recommended we spend our budget? At what point have I automated my own employability?

  6. States that the risk assessments it cites are the AI developers' own published assessments, not the submitter's own claims.

    These are not claims I am making. They are the developers' own published assessments, and they are the reason I am writing.

Recommendations

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

  1. Hear independent experts on catastrophic and systemic AI risk, including AI-enabled biological and cyber capability and loss of control, and address it in the Committee’s report.

    That the Committee hear directly from independent experts on catastrophic and systemic AI risk, including AI-enabled biological and cyber capability and loss of control, and address it in its report.

  2. Push for international agreement on verifying and pacing frontier AI development.

    Push for international agreement (items m, e) on verifying and pacing frontier AI development.

  3. Have Australia lead international pacing efforts and press for agreements with verification, beyond commitments.

    That effort needs countries willing to lead it, not just join it. Australia has done this before, we set terms for global platforms on social media age limits and others followed. We should do it again here and should press for any agreement to include verification, not just commitments.

  4. Resource the AI Safety Institute to require access to frontier systems deployed in Australia, compel information from developers and deployers, and give its assessments consequences.

    Empower the Australian AI Safety Institute (item m) to do its job. Give it the power to require access to frontier systems deployed in Australia, to compel information from developers and deployers, and to have its assessments carry consequences. Then resource it to the level that job requires.

  5. Introduce mandatory incident reporting: a reporting regime for serious AI incidents.

    Mandatory incident reporting (items i, l). A reporting regime for serious AI incidents. We require aviation incident reports so that near-misses are visible before a crash, not after.

  6. Legislate adaptable principles and thresholds with statutory annual review.

    Legislate for adaptability (item i). Principles and thresholds capable of being updated rather than a fixed schedule of applications, with statutory annual review.

  7. Produce public scenario analysis of high-automation futures, by Treasury or a body the Committee nominates.

    Public scenario analysis of high-automation futures (item g), produced by Treasury or a body the Committee nominates.

  8. Set the terms for frontier AI development built on Australian soil, including in planned data centres, before the builds start.

    And where this is being built on Australian soil, in the data centres now being planned here, I would like us to have set the terms before the builds start.

Read the original

Open in the interactive guide