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 424

Name withheld

Submitter group: Individuals.

Position labels are not shown for individual submitters.

An AI model identified the submitter group and chose the extracts. Counts do not measure importance or merit.

Open in the interactive guide

Extracts from the submission, in the submitter’s own words.

Extracts

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. Australia cannot and should not attempt to lead the frontier, and we should not pretend otherwise. But we hold two things that matter enormously in the next decade. These are land and firmed energy at scale, and a stable democratic state with a functioning regulatory and judicial system.

  2. Treat hosting a material share of frontier AI training compute as a national strategic objective, with Australian safety and transparency standards attached as a condition of access.

  3. The International AI Safety Report 2026, produced by over 100 independent experts and panellists nominated from more than 30 countries, describes an evaluation gap: pre-deployment testing does not reliably predict how systems behave once released.

  4. Build genuine sovereign capability in evaluation and in the open-weight ecosystem, rather than attempting to build a frontier laboratory.

  5. The money in consumer AI runs toward engagement, because time in app, retention and daily actives are what the business models measure. But I believe it may be possible to build the opposite.

  6. The question is whether we treat that as a commercial opportunity to be maximised or as leverage to be spent on conditions. I would urge the Committee toward the second.

Extracts from recommendations

Extracts chosen by an AI model, in the submitter’s own words. Listed in the same order as on its page in the interactive guide.

  1. That the Government pursue, as a defined strategic objective, the locating of a material share of at least one philosophically aligned frontier developer's training compute in Australia. That should include meaningful training, not inference or general cloud capacity.

  2. That any access we provide be conditional: adherence to Australian transparency, evaluation and incident-reporting requirements; guaranteed government access for independent evaluation (we need to also build up this capability); and commitments on the security of the facility itself.

  3. That the Committee coordinate with the Senate inquiry into AI and data centres, so energy, water and community impacts are settled as part of the same decision rather than discovered afterwards.

  4. Australia should not try to build a frontier laboratory.

  5. A national evaluation capability that is publicly funded, technically serious, and has contractual access to the models it assesses is a small investment against the scale of what is being decided.

  6. But this ecosystem will exist whether or not we participate, and there is a public interest role in it. Models and datasets for Australian languages, contexts and public services. It’s also more affordable! Tools for small business and regional users the commercial market will not serve. And the technical depth that comes from building and fine-tuning rather than only buying finished products, which would further the evaluation objective as well.

  7. providers of AI used for companionship, emotional support or anything close to it be required to disclose publicly what their systems are optimised for, and which engagement metrics they track internally.

  8. Australia fund independent research into which design choices actually drive better and worse outcomes in society, so the evidence base is not owned by the companies with a commercial interest in the answer.

  9. That we help establish guardrails that distinguish between systems designed to retain users and systems designed to help them get off the app.

Read the original

Open in the interactive guide