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.

Require independent safety testing of frontier AI systems before deployment

7 submissions. Each submission counts once for this shared ask. Counts do not measure importance or merit.

A shared ask groups requests for substantially the same action from three or more submissions when the group was created. Groups remain listed if withdrawals reduce their counts. AI models summarised recommendations, grouped them and checked the groups. The label summarises the requests; it is not a quote. A shared ask does not mean agreement on anything else.

Theme: Frontier AI governance

Recommendations from submissions

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. Extracts are chosen by an AI model. Submissions are listed in submission number order; recommendations follow their order on each submission’s page.

  1. Good Ancestors Submission 257

    1. Secure oversight of AI development in Australia, including pre-deployment testing of Australian-trained models by the AI Safety Institute, disclosure of capabilities and risks, and reporting of critical safety incidents.

      Oversight over AI development in Australia, including pre-deployment testing of models trained in Australia by the Australian AI Safety Institute, disclosure of model capabilities and risks, and reporting of critical safety incidents to Government.

  2. Mrs Elizabeth Morrish Submission 390

    1. 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.

  3. Ms Alana Kane Submission 392

    1. Require independent safety evaluation, with appropriate confidentiality protections for commercially sensitive information, rather than testing solely by developers.

      Testing should not be solely conducted by the companies developing the systems. There should be independent evaluation, with appropriate confidentiality protections for commercially sensitive information.

  4. Name withheld Submission 393

    1. For highly capable frontier systems, I believe this framework should include independent safety evaluation once defined capability thresholds are reached. Areas warranting particular scrutiny could include sophisticated cyber capabilities, dangerous biological assistance, significant autonomous operation, autonomous AI research, evasion of human oversight and autonomous replication or persistence.

  5. Miss Michele Bray Submission 441

    1. Legislate requirements for independent external verification of frontier AI systems before and after deployment.

      That the Committee recommend legislated requirements for independent, external verification of frontier AI systems, before and after deployment, rather than relying on voluntary industry commitments.

  6. Name withheld Submission 477

    1. Pre-deployment safety assessment for models exceeding defined capability thresholds

  7. Mr Craig Weniton Submission 502

    1. Require secure pre-deployment evaluation access under statutory confidentiality protections for model weights, trade secrets and security-sensitive findings.

      Require secure pre-deployment evaluation access under statutory confidentiality protections for model weights, trade secrets and security-sensitive findings.

Open in the interactive guide