End Up Here

South Australia · inquiry notes

Make it
possible.

Artificial intelligence should be allowed to go as far as it can—while people remain able to understand it, challenge it and call on the law when access or power becomes dangerous.

Current statusInquiry notes · no formal submission lodged

01 / The inquiry

What is known
so far.

South Australia’s Royal Commission into Artificial Intelligence commenced on 1 October 2026. Dr Iain Ross AO is the Chair. The appointed panel also includes Kate Pounder and Professor Christopher Manning. The Commission is due to report to the Governor by 1 July 2027.

The formal Terms of Reference ask the panel to examine AI’s economic and societal opportunities and implications in South Australia. They cover productivity, innovation and economic growth; social benefits and risks; policy, legislative and regulatory frameworks; employment and workplace change; education; streamlining regulation for businesses; and public services, including health.

The panel must also consider human-centred innovation, safeguards and accountability; trust, civic life and democracy; data sovereignty and South Australia’s strategic autonomy; other jurisdictions’ experience; safe practical pilots; views from community and stakeholder groups; ways to scale existing initiatives; and the costs and benefits of recommendations. The official Terms of Reference state that data centres and matters primarily within Commonwealth responsibility—such as national security and cyber security, financial-market stability, critical infrastructure, and intellectual property and copyright—are outside the Commission’s scope.

This page is a working record of my questions and proposals, informed by the published Terms of Reference. It is not a claim that a formal submission has been lodged.

Royal Commission website and commissioner details ↗Official Terms of Reference summary ↗Formal Terms of Reference (PDF) ↗

Illustration and scene

A practical remote Australian settlement appears in an editorial cartoon.
The wider ideaAI-generated editorial cartoon
A small South Australian settlement sits at the edge of red country beneath a wide sky.
Everyday lifeAI-generated illustrative scene, not documentary photography
02 / Proposal one

Free up
the future.

South Australia should encourage AI to develop and be used widely. The state should make it easier for people, small organisations, researchers, schools and public services to learn what these systems can do and apply them to real problems.

  • Prefer experimentation, access and practical education over blanket bans.
  • Use AI inside government where it can improve services, discovery and productivity.
  • Measure outcomes in public: what worked, for whom, at what cost, and with what risks.
  • Build local capability so South Australia is not only a consumer of systems made elsewhere.
03 / Proposal two

Put a human
at the gate.

AI will increasingly hold, infer or generate information that matters. When a person seeks a warrant or other legal authority to access sensitive AI-held information, the decision should be made through a clear human legal process—not by an automated system or an informal technical shortcut.

  • Create a fast, specialised legal and policy faculty for AI, privacy, evidence and security.
  • Train lawyers, judges, investigators and public officials to make timely, defensible calls.
  • Require necessity, proportionality, purpose limitation, authorisation and an audit trail for sensitive access.
  • Keep a real human accountable for high-impact decisions, with review and appeal pathways.

04 / The balance

Capability and
constitutional care.

The mistake would be to treat safety as a reason to stop learning. South Australia should become unusually capable at using AI—and unusually good at protecting people when systems touch liberty, identity, health, money, employment or public power.

The answer is not human involvement everywhere in the same way. It is graduated governance: light-touch support for low-risk uses, strong transparency for public systems, and rapid human legal intervention where access, surveillance or coercive power is involved.

05 / Questions for the Commission

What should
we ask?

Who gets to use AI? Can ordinary South Australians access useful systems, training and advice—not just large organisations?

Who is accountable? When AI informs a decision, can a person explain it, challenge it and correct the record?

What requires a warrant? What legal threshold should apply when sensitive AI-generated, inferred or stored information is sought?

Who is ready? Does South Australia have the specialised legal, technical and civic capability to respond at the speed of the technology?

06 / Review trigger

Next consultation
review.

This is a public record of Martin’s current questions and proposals, not a lodged submission. Review it when the Commission publishes its issues papers, expected in late 2026 and early 2027; decide separately whether to prepare and lodge a formal submission.

Last updated 4 October 2026 · No final submission is represented here as lodged until there is a confirmed record.

Discuss the submission