Submission 202
AQ Intelligence (AQI), Guardrail Technologies Australia (GTA) and First Nations Strategic Alliance (FNSA)
Submitter group: Companies.
An AI model identified the submitter group and wrote the summary, key points and recommendation summaries. Counts do not measure importance or merit.
Summary
Read the original submission for the full text.
Joint submission from AQ Intelligence, Guardrail Technologies Australia (an AI guardrail and assurance venture in which AQI and FNSA hold shares) and the First Nations Strategic Alliance. Supports technology-neutral, principles-based governance, argues it depends on verifiable implementation, and recommends an assurance pathway, sovereign guardrail capability, proportionate obligations, First Nations data principles, workforce measures and consent-based copyright.
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.
Argues that a principles-based regime only works if organisations can operationalise it, and that governance capability is enabling infrastructure for adoption.
A principles-based regime only works if organisations can operationalise it. Governance capability is enabling infrastructure, not a brake on adoption.
Argues that the main gap is verifiable implementation of existing principles rather than a shortage of principles.
Today, most organisations cannot independently demonstrate that an AI system meets even the GfAA’s six essential practices. That is the real gap — not a shortage of principles, but a shortage of verifiable implementation.
A program corrected obvious text-recognition errors in this quote for display.
Asks the Committee to recognise the guardrail, assurance and monitoring layer between imported models and Australian users as a form of sovereign capability.
We ask the Committee to recognise a distinct and achievable form of sovereign capability: the guardrail, assurance and monitoring layer that sits between imported models and Australian users.
Proposes proportionality within technology-neutral standards, focusing effort on high-risk deployments rather than a prescriptive tiered Act.
This is fully compatible with the National AI Standards — it is not a call for a prescriptive tiered Act, but a lens for deciding where the six essential practices should bite hardest. It concentrates effort where harm is most likely and most severe, while leaving low-risk innovation free to flourish.
A program corrected obvious text-recognition errors in this quote for display.
Recommends treating Indigenous data governance as a benchmark for public-sector and high-risk AI.
We therefore encourage the Committee to treat Indigenous data governance not as a niche carve-out but as a benchmark for public-sector and high-risk AI.
A program corrected obvious text-recognition errors in this quote for display.
Argues that barriers to adoption are governance barriers, including assurance gaps, liability uncertainty and limited skills, rather than technology barriers.
The common thread is that these are governance barriers, not technology barriers. Each is addressable by infrastructure and policy that make the safe path the easy path: recognised standards, assurance/certification, clear liability allocation and onshore governed deployment options.
Recommendations
Recommendations are shown beside the submission’s own words. Listed in the same order as on its page in the interactive guide.
Establish an independent AI assurance and certification pathway for organisations to demonstrate they meet the Guidance for AI Adoption's six essential practices, aligned with international standards (NIST AI RMF, ISO/IEC 42001).
Establish an independent AI assurance and certification pathway that lets organisations demonstrate they meet the Guidance for AI Adoption’s six essential practices — so that a principles-based regime rests on verifiable evidence, not self-declaration. Align it with international standards (NIST AI RMF, ISO/IEC 42001) for interoperability.
Apply proportionality within the technology-neutral framework: minimal burden for low-risk uses, and the strongest obligations and independent assurance for high-risk uses affecting health, liberty, essential services or safety.
Apply proportionality within the technology-neutral framework. Let expected practice scale with the stakes of the use case: minimal burden for low-risk uses; the strongest obligations and independent assurance for high-risk uses affecting health, liberty, essential services or safety.
Treat sovereign AI-assurance and AI-security capability as national infrastructure, extending sovereignty beyond compute and data centres to the guardrail, monitoring, assurance and security layer, onshore data handling and an AI-governance skills base.
Treat sovereign AI-assurance and AI-security capability as national infrastructure. Extend the National AI Plan’s sovereignty logic beyond compute and data centres to the guardrail, monitoring, assurance and security layer, secure onshore data handling, and a home-grown AI-governance skills base.
Clarify accountability across the AI supply chain by allocating responsibility between developer, deployer, integrator and professional user.
Clarify accountability across the AI supply chain, allocating responsibility between developer, deployer, integrator and professional user to remove the liability uncertainty that is freezing adoption.
Adopt CARE and Maiam nayri Wingara First Nations data-sovereignty principles in public-sector and high-risk AI, and support First Nations ownership models in AI ventures.
Adopt First Nations data-sovereignty principles (CARE and Maiam nayri Wingara) in public-sector and high-risk AI, including consent, benefit-sharing and community authority to control, and support First Nations ownership models in AI ventures.
Invest in an AI-governance workforce and inclusive transition, building assurance and oversight skills and protecting entry-level pathways, care workers, people with disability and First Nations communities.
Invest in an AI-governance workforce and inclusive transition, building assurance/oversight skills and protecting entry-level pathways, care workers, people with disability and First Nations communities.
Legislate the copyright settlement on a consent basis: Australian creative works not used to train AI without creator consent, with training-data transparency, workable licensing, and clarity on ownership, provenance and liability for AI outputs.
Legislate the copyright settlement on a consent basis. Give effect to the Government’s stated position that Australian creative works not be used to train AI without creator consent, backed by training-data transparency and workable licensing, with clarity on ownership, provenance and liability for AI outputs.
Default to onshore, governed deployment for high-sensitivity data (health, disability, legal and First Nations data), with strict data-egress controls.
Default to onshore, governed deployment for high-sensitivity data (health, disability, legal and First Nations data), with strict data-egress controls.
Treat AI security and AI governance as a single national capability.
Australia should treat AI security and AI governance as a single national capability.
A program corrected obvious text-recognition errors in this quote for display.
Preserve meaningful human oversight roles in high-stakes settings rather than allow full automation.
In high-stakes settings, policy should preserve meaningful human oversight roles rather than allow full automation — protecting both safety and skilled employment.
Provide people affected by AI decisions with transparency, a right to human review and effective redress.
People affected by AI decisions need transparency, a right to human review, and effective redress.
A program corrected obvious text-recognition errors in this quote for display.
Support provenance and content-authentication measures and clear offences for malicious synthetic media, balanced against legitimate uses.
We support provenance and content-authentication measures (e.g. watermarking/credentialing standards) and clear offences for malicious synthetic media, balanced against legitimate uses.
Align Australian rules with mature international frameworks (NIST AI RMF, ISO/IEC 42001) while retaining sovereignty over enforcement.
Interoperate internationally. Alignment with mature frameworks (the NIST AI Risk Management Framework, and standards such as ISO/IEC 42001) reduces compliance cost, supports Australian exporters and eases the burden of trading into the EU regime, while retaining Australian sovereignty over enforcement.
A program corrected obvious text-recognition errors in this quote for display.
Build security-by-design, continuous monitoring and incident response into any high-risk AI deployment.
Security-by-design, continuous monitoring and incident response must be built into any high-risk deployment.
Embed free, prior and informed consent for First Nations data in AI training and deployment, with community authority over secondary use.
Embed free, prior and informed consent for the use of First Nations data in AI training and deployment, with community authority to control secondary use.
A program corrected obvious text-recognition errors in this quote for display.
Require benefit-sharing where AI is built on or applied to First Nations data or communities, consistent with collective-benefit principles.
Require benefit-sharing where AI is built on or applied to First Nations data or communities, consistent with collective-benefit principles.
A program corrected obvious text-recognition errors in this quote for display.
Adopt Indigenous data-governance standards in public-sector AI procurement and any national AI assurance scheme.
Adopt Indigenous data-governance standards in public-sector AI procurement and in any national AI assurance scheme.
A program corrected obvious text-recognition errors in this quote for display.
Treat Indigenous data governance as a benchmark for public-sector and high-risk AI.
We therefore encourage the Committee to treat Indigenous data governance not as a niche carve-out but as a benchmark for public-sector and high-risk AI.
A program corrected obvious text-recognition errors in this quote for display.
Consider illustrative minimal-risk practices of voluntary best practice and AI disclosure without new mandatory burdens; elevated risk would add requested human review, records, bias testing and recognised standards.
Tier illustrative obligations Minimal risk Voluntary best practice; transparency that Alis in use. No new e.g. productivity tools, mandatory burden. drafting aids Elevated risk Disclosure, human review on request, record-keeping, bias testing, and e.g. customer-facing adherence to a recognised standard. decisions, workplace tools
Consider high-risk oversight, monitoring, certification, audit trails, incident reporting, data-egress controls, accountability and independent assurance; the proposed illustrative tiers prohibit unacceptable manipulative or rights-violating uses.
High risk Strongest obligations, and the case for independent assurance: e.g. health, disability care, meaningful human oversight, continuous monitoring, certification, audit Sey SEES trail, incident reporting, data-egress controls, and clear accountability. safe! Unacceptable risk Prohibited. e.g. manipulative or rights- violating uses
Read the original
In some quotes a program corrected obvious text-recognition errors for display.